From 6ba4fd47b4fdee3746460b5f4fd697a982c693e8 Mon Sep 17 00:00:00 2001
From: Joris Van den Bossche <jorisvandenbossche@gmail.com>
Date: Tue, 16 Apr 2024 17:57:12 +0200
Subject: [PATCH 1/8] Add test data covering different native (geoarrow-based)
 encodings

---
 .../data-linestring-encoding_native.parquet   | Bin 0 -> 2017 bytes
 .../data-linestring-encoding_wkb.parquet      | Bin 0 -> 1468 bytes
 test_data/data-linestring.wkt                 |   3 +
 ...ta-multilinestring-encoding_native.parquet | Bin 0 -> 2211 bytes
 .../data-multilinestring-encoding_wkb.parquet | Bin 0 -> 1802 bytes
 test_data/data-multilinestring.wkt            |   4 +
 .../data-multipoint-encoding_native.parquet   | Bin 0 -> 2022 bytes
 .../data-multipoint-encoding_wkb.parquet      | Bin 0 -> 1616 bytes
 test_data/data-multipoint.wkt                 |   4 +
 .../data-multipolygon-encoding_native.parquet | Bin 0 -> 2411 bytes
 .../data-multipolygon-encoding_wkb.parquet    | Bin 0 -> 2258 bytes
 test_data/data-multipolygon.wkt               |   5 +
 test_data/data-point-encoding_native.parquet  | Bin 0 -> 1815 bytes
 test_data/data-point-encoding_wkb.parquet     | Bin 0 -> 1375 bytes
 test_data/data-point.wkt                      |   3 +
 .../data-polygon-encoding_native.parquet      | Bin 0 -> 2190 bytes
 test_data/data-polygon-encoding_wkb.parquet   | Bin 0 -> 1857 bytes
 test_data/data-polygon.wkt                    |   4 +
 test_data/generate_test_data.py               | 195 ++++++++++++++++++
 19 files changed, 218 insertions(+)
 create mode 100644 test_data/data-linestring-encoding_native.parquet
 create mode 100644 test_data/data-linestring-encoding_wkb.parquet
 create mode 100644 test_data/data-linestring.wkt
 create mode 100644 test_data/data-multilinestring-encoding_native.parquet
 create mode 100644 test_data/data-multilinestring-encoding_wkb.parquet
 create mode 100644 test_data/data-multilinestring.wkt
 create mode 100644 test_data/data-multipoint-encoding_native.parquet
 create mode 100644 test_data/data-multipoint-encoding_wkb.parquet
 create mode 100644 test_data/data-multipoint.wkt
 create mode 100644 test_data/data-multipolygon-encoding_native.parquet
 create mode 100644 test_data/data-multipolygon-encoding_wkb.parquet
 create mode 100644 test_data/data-multipolygon.wkt
 create mode 100644 test_data/data-point-encoding_native.parquet
 create mode 100644 test_data/data-point-encoding_wkb.parquet
 create mode 100644 test_data/data-point.wkt
 create mode 100644 test_data/data-polygon-encoding_native.parquet
 create mode 100644 test_data/data-polygon-encoding_wkb.parquet
 create mode 100644 test_data/data-polygon.wkt
 create mode 100644 test_data/generate_test_data.py

diff --git a/test_data/data-linestring-encoding_native.parquet b/test_data/data-linestring-encoding_native.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..d4853894ddbcdb21afea8c2d2d3f14f667f9a885
GIT binary patch
literal 2017
zcmc&#&2Q6K5Px=@G$0NI?Xw&yhbRhG51}LlBZ?}NNm3e@q#r>vX;ug&sW<D$`M`0S
z80ekzp7zLb|BO9y<cK(N<Zs}_%yZm=R0K$TJlXHf_nY6$Sl)wro+~`XGgUm;6~>k~
z*kVbGlQVvWt69`9^B9jMwRot};-Rg|R?d;4_CiX=U6G`StiO?xTuB%#ww%x+-I2p(
zRZ(9_>T5}_=<)A~o1knMml<x$crb>1j%TZwAc4>ri%QDfLUd_`WeTXqRW6`$n(J8}
zKLrwMLLixfDz7X+l?!K5w7XoOYa+1-?R_RkGED!VB>BYvP$k~8Mgz-peZ{dS-jd~5
z18e9>a|9Vz-%I%DZzK9Y5rp~8w|*PUjDUH!5QhEXFki*L3e*1t=6_52JDQR(pOE@u
zanFO9cy2H)HtGIt4cc76_Fk36%rHaNPD{;Y9;?#48j2>$j-!BXVe!fD;#H<e^Tp{i
zRY@ch8$Wdtf5TMhzrI82zfo1u<Lb-kAB-ikS_H>mdm%kbA6xFk9u3oXQtA9^el?fQ
zrqW~A9<*KmF@E#Yfl!4X3uy!C1U;UmPvyi|!|v$F9`-1fV-Kx~=Msl5q1j{4A6uyY
zGhIdfgFyf2<m7~Dm!MwXZQKbaNc%r}oz|>Z0)TydY26373B?}V1L#4iCuGIqXZjfl
zH&6A|kVmPAB6t&E2Iz;U2!ba-nKr25TO76;x4|a1GO>>a2Yz?WIqulzEwg0KtNqPs
ztz=Hj()O%7aJ<9jw$WVMIqp;ryufA~{nx!#v+i3Dgk8V0HqUn&##6ID>RJA0`dcR6
z(P|iIQy+H*U3=FCyR+p?Tg{yu#-Fx)5#uc8!FtYW79EP!@V9I2hT%NCQTOp2nofTG
zuV#)i51{P|&{z+Eojs~268p1<#0$_tQLq#MTeOLK5dKB}!o6IhR*XV_$Q_a;HUpP1
zpdLYkwqjUc0IJ*3meX%FsivqBeIfhuA;lBIx!6TMA_>8gDS$<$H5V;!BVkfg(N6TM
zLjA+FA6OCzWdRb01D7C)2*~11fJ49}5u!*P<!ABG<(zn%r68WT=aYD1KCo9aiib>(
zAP1$S>jV-a0GO5<x7{aG%e&SckFT{|cQi{aKDJceSk0~G5*tazf`=8OjP2pi{Ud>v
Bj<^5-

literal 0
HcmV?d00001

diff --git a/test_data/data-linestring-encoding_wkb.parquet b/test_data/data-linestring-encoding_wkb.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..853ac9ee231e6c46f9f9cc3b9df90acd7dc529d4
GIT binary patch
literal 1468
zcmc&!&u<bz6n+aU0aFv&IGf#s10j|4KoL-y(liaTP-Q9oMHOg`Hn6PQtt{ZOEybn>
z{{auigMW{SNsk_lC;tvlzL{kkYkD!!__8~1=KJ1zGw++t?$COch$K$tYv^DOAi4@u
z6Q-U3nIlpP`6vmIP)rW{j2!l3C0MutMK%%-;3x`0Fp&Hz#E7WmArw{Qpg!o4fFw$n
zg7ig5R+8aAsjINl00{sNQFH*f{vePSeizk+0)hf#f-ozHYsJWIm|qFQtPJHl@^ZNt
z@p<&D%(WJmV4*NgW;C<4B6l$VoA>eZ6fYO$2v7CGuL<z&R*cL{1qhqQz&2cG!d3VY
zkS+tscVhB=F!t9rzey!!H4k!Vq#qe}n}{gey#T)pATv0-IuuBv5>rzD^vaU|T@LA!
zD2d6i^ijMEprqs=-iiEHI8U4!&d3_{6D#pVHl0mp5~+A%=va2!nH=JWdt!6Szd^=&
z=m;~;6IVD}7=3-vwfZKj)!a~7tj{01lc9n5vqa768+)$9!k%ATT!1`7>-FvCig!=)
z_uzeIIjT}hckyMmLFq%Pn)DH+CM}s9Q!(Hde$2r2E561lT%++R=QT>ltT1|B1U;o?
zHqcFc)&4<qnVKxVGP1gMW1{DJryWaM(n{J{?RfoUv!so*Qgy7`J$JuV&9`z}r=3c}
z^<t`+UvLjv^@*{^<BmJIvuvlCf1@1_Ok;A5U(#@ogJvFM>cft$TiX`38X32wSz}YP
zyS+{>+v`?vj|cU<{`e{KYDar@hOvz&i1VCQD_eNU`(v%BydNeU-72Gq%Q#tI=OUCC
z$n^1z_0Rg9o|o5P{4N)s4JbC&+o$T7i{p`Ax=N4uL}w%4&%?Qt`?<o1>5B`@*QvKP
v3&4tKAjgMo=k<x<F6zVKV%u>B<M{L|jA!%dOgf{aVgUHS=ODli{<nVu238T7

literal 0
HcmV?d00001

diff --git a/test_data/data-linestring.wkt b/test_data/data-linestring.wkt
new file mode 100644
index 0000000..26f6fa4
--- /dev/null
+++ b/test_data/data-linestring.wkt
@@ -0,0 +1,3 @@
+"col","geometry"
+0,"LINESTRING (30 10, 10 30, 40 40)"
+1,"LINESTRING EMPTY"
diff --git a/test_data/data-multilinestring-encoding_native.parquet b/test_data/data-multilinestring-encoding_native.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..fb4780abdb7b6b28bb2c3d50fe8daadac94060fc
GIT binary patch
literal 2211
zcmc&$L2nvI6rNpRO{_$+D$jP2I3Pr}mP!Z^Zxlyv#4Mz+6daqhfQhOImd$4EU7+k@
zSVz9)*i+R*ROL^##~yR)p+BLAzL^0WOHrDriZX(onfJc$&HLWG#Xcmh0*>N5&Ni@w
z(H96Q2;C6&1;mHAoH4@NSjhrr3%*iPiLX3Ld<j7^+Ifc+yqg&cmIN`9LFyYJiK7V(
zNm~geVvb!b%98x203Y>GjektwL>4kfSVZvHf(aqmm+)=_$C;88LNOuwpcLDPqRfqW
zrcwe3jNK9ta=4hq@>&h|R7xu~nUXB-ES0@&1>Be27(8P;zipx@EVd(}AjXDYiAa4F
zOX5u~?Rng`#-0^S!>H@{!G`5po;3=tY=i}9m*tnD{Ewji9Z_FLK>Y_N@?q*TO#OpT
ztUhxZLF%v9K2Cl3mioUo)K?q-SAF8Oh7`#+pIcrZN1+*RimbzsCMxUez#^6!`%q;q
zs<7LD#sS*ug8s`5VK;9gMOZB75Xn*^nb`lNZT-0$>vLbN{3a$#YFz#!wvA9Ct3=@4
zQhpI$q|dCW?~F(3`>AvxSIFhl*;INmb-eyGd<s7;Gmp7eZzA<_s}GD9>9w71tdTh$
zIwPCL@@8(}xX#G(gDEuuZ-sptgcA!ePtr{aY5?dl2mN?)ae<WUq}6)VxzC$W`B}P6
zN1l`kAxH3~<2yojiEfi`39(7pW;R{F;-6FF!?nN3sIZ5@mdtMuGAF=;nFWEFkP3Z}
zCVcf#uXB%_QdiA)hF&`~i|*OLG42{=W6?NyIBS*--zeATrsoF7-MZE-9-Iwo?SNzI
zz~2jc-BxHlX1J3<aZwm_v~%NRY+K<K|6K#l=yf!}v?c@3bRIdx8FbtJu<H8#j^>(o
zS|Q9M!<pNLH*^QZf;+6inRi;6S^WX<`oK6X6b7}{<ol{S>$MN^kk_mi)`!o!g*{L~
zYY0W7W#ILorG<@N(b=9A0qOyRmFk$T7^!-B9<`DBIWYI;OFnropO(7108Q<4$QS0}
zIcZY<*Y=v<Vn>f@XnsDu3Z0!_qegZ=da>87=lh)l*JRg6lRV=1gA)jKj{++!bFPOi
zsgdi^HM$DV0805bB_DBp#N_u5hhLI-(^M&m*Nd&#vS!o%gCvjX%*SOspu8jAk->`0
z^{^%9vHC#Yus)pAd;qcvN;6!KZ;+Cj^rt_~tl*Y8ncV76r{j5Q`3k2BS}vc<C-##F
P@eleP{>&c2zbF3!9Wlf}

literal 0
HcmV?d00001

diff --git a/test_data/data-multilinestring-encoding_wkb.parquet b/test_data/data-multilinestring-encoding_wkb.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..1357eb2e416c338e6aae2e6e0eab8165ae496cd2
GIT binary patch
literal 1802
zcmds2UvCmY5TAwPa&6kAm>xM29tc5F6IwYeO-pLS9<=3Z|3MWh#x%X-q}TFKIEvWx
z)tL0vM?Zs~!uSb%Fh2V1hj4a}Bi8uhgNboqcW&nQ+u5DjIXI+cRTkuwoZOW~8SVfm
z0JCJBfGfVSmO-AEl_YZJaEK8_@(@KDQy>HjzwshE9cN^b`22B5d?OKA2xmZ?3oCxJ
z)0BN7G4z$-H*uJd{v>cXi6;ly2RO~40zmttU;aY8T)vTc4c#PwA7P#`{17C<Ms8{v
z;#moT3KV7)sgS!N56Kpx*)7=E09c+~2};4JP{;+=K$vn^hT#){00|Bm0L$4Ztj2Rw
zJT^{Deyz|7EZh(hK!oi}+|3!0Nj%5ZL@$qRPkVSF#}^7N8|T89=;@8I?P<^N1@ipy
zoHE4^GmdoYO5REld1ibiQp4)BET=c(_wrc`eHRk%0*Uwj$iJM>?^+?eo&iPdn`c(*
zR2Hx|?)rGcfjomxlRP0Lh9lwRU+v4p|5)MB2Pq^bq|hgE9$$~7;>XTW-jJ)<rPb@(
zopvl6jj2oOQYw~=#=1Sbb=n)9;0r%!amed4hPvK9uDFU#{LHZ0W~Xkq8!T4!k;b?l
ze&UR}7W$vX_6AMI-nHA-k<(+`=hxTQpv=&6`JkG0?_c@uzt1Gy&?u!P984Zjx=ghO
z-J-NX^9}CPknj^eWX{bAzRO8yqi`0-bxMa6*P-S(YDx=iqI)>BcCETf&lptf+x1pu
zWTu;!hOMvYdHvPy`Q~6RulM!*&d_W%ou|h;nd9{SrBSRnE@ubVuQ|2j^2j>kymKS{
zN;RsP*ZO&<VU2F#D?08`t7ec>?iwxAKCr1>NjZ7l9yat=y=kP?X1$1eJSb<(?I(z9
zhJK-{MzP#|yxknsD*Gus$DlUaslPZ@*YMPN9Q24yJn2G-l@bk&6$^zY?LKmw=U+=#
zE{$p>WwtN46LG8?FItIeB^Gp>Qq!%I3vwPU>6}#LmUk6suu^YhU3i^`Y&}x=){Kb}
vC-M-KX!o@Da$q@&X1BX|+Us?O(Q#Ww)yz_ADHTpe0PqX<@W=HK{{?>lm%=Nf

literal 0
HcmV?d00001

diff --git a/test_data/data-multilinestring.wkt b/test_data/data-multilinestring.wkt
new file mode 100644
index 0000000..17294fa
--- /dev/null
+++ b/test_data/data-multilinestring.wkt
@@ -0,0 +1,4 @@
+"col","geometry"
+0,"MULTILINESTRING ((30 10, 10 30, 40 40))"
+1,"MULTILINESTRING ((10 10, 20 20, 10 40), (40 40, 30 30, 40 20, 30 10))"
+2,"MULTILINESTRING EMPTY"
diff --git a/test_data/data-multipoint-encoding_native.parquet b/test_data/data-multipoint-encoding_native.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..ef9791ea9faff1a1bbd88f4a9aba50174a96337f
GIT binary patch
literal 2022
zcmc&#&2H0J6ux#6H>!%%8F4K~$|8z_#UgZ^sz$UbC^reyIFq(XMkHwwLWyI7D?2fX
z^JAd<S?~xv0Z)Kfvg8q1B9^ll&1#l&&UKQ4gailySMmKhKi@syIp-=h*2vK))o8X#
zB}#rJL?PsourH8sh>MjHdYLL&u-Sx9N-FWCMu{&>h)k}2LJHdT3>r&<7|9Uzosgu_
zgifT*gc31_4i#ldek;IFt*P;^0ZwEgazaG{(<Us0z&THMsx;1-qzH)#(Odc0T9jli
z#WSTmn84V{gF~9CS-P=e6KYENC7VoM7OyUteJBOHU-EcWBpJc^g(ODa6ZKgvN!LSC
z<#^v3x|Tl*qK-ZG*DS|!t${Dhf#`<(pCJDusDDS)=MkX##kYPL)eNUv0;;$2A@p~k
z`Y!%is=)_TU$3eEV!M13)x=YsDAL$`Xt^C4g)(0edG8QiR95MxB9*H9aE~5Uc<P}g
zz-?Io^jp6|6k)zt10qX_WMcojcH~d!4E4uXNPZKOB{eR;id`lokyRq_1eCvoC+Vp*
z8r#D`dM}mEZRNJKbT*atMz-4-1^40YPh75AJ{Z)7(lK~ENw4I1um<L^XAk;_<xU*m
z_J;PrN0+79eLwIlP~S}-qQ1uU_s-AHiL${OjpNo{*g?w6*l##ytjHLv!-wM!#&%et
z&wgjD&x(C6D-=G`&ro=KrLQi}Amu3hY%n%s;0G%Qft9fmPV5lA!r-)ZgV}g2kL{k@
z49sn3+O>`AM$wp8&u&i+i^kX}R%WK__$Td(-rhc%cFRpaw5fpqu7BEY1lFJ2?yS2#
z&vje+W8-Ytw}OxK*A2Miw55Yh!|S@HeQYzk+is3~2hOa!?M!;%T^E?mUCnA29K>n`
zm7`v3u63G)X{V)YhLhWU(AF^WklHcdK|UBeuJL?A9%gs&%t?nyfmASdfD_LL@?XTy
zb5^QjDZHgW(@xN`f>Y>^HDXPipg1r%lzFi0bq9^daEANqm~Y08&H(8K?n*qU0>6N7
zQOHeMLC0M%UBuf*=Q?i$#>}uz0CwSY$U(yR`4-*7(8_%b38<bdLIe2)9=Lo0uWoRV
z@(P4+uw{T5)S0J_Hzt^HzGx6KVK;9QCFONSk0zFX&Gfu$ozZAGOD!H*DyMI0TUug2
PNl5r3doe<4@DKk#ehQGo

literal 0
HcmV?d00001

diff --git a/test_data/data-multipoint-encoding_wkb.parquet b/test_data/data-multipoint-encoding_wkb.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..9fd760557604428b15690059d31ea5e8c7903f1f
GIT binary patch
literal 1616
zcmdT_%}(1u5FRJS3KdAGVl7(=Rb-_Rhmgi;f(Sw=>m&v9Pbx^DLMV3JV4A-shL9o-
zeS<zg4?I8*eU=`2sCuY+>$OkN*>yr`#ifT<>c~4g`+c*syWeh<9a_?fNRlMJK_mkA
z0Th5qVMzer0>hpFnI=jcbEdFKVI?$Tl+Z8*WSG4JMfP5lB}+me7=_qZAxuOy1JabL
z1nqW%1Y}A6B;Xg@iG}{8uo}mm0|@}SDeM4nd>tUy!ssk-gy{ze3!*T_+PEMt=f*>!
zXzoZ*Cq&N7Xb_#$#6oU{Kw*-=a)JP)F-T*eR+HmA3mvoHRhWm_BZ1Z7wMLo&L~|?~
zppX-Of`uH%VEHzf9)juH;PMcR<`mqrKa2Tabo?%b$;42{P~B-a9j`m!Vtx?ikHOey
zG4?hX{--;<>q69}3@B31K5?2=BH|4_4Df3QGJ(;}#zB_Uup0lXuOs$fNG89NWGNPs
z-;YfLsBtBTcdNV>F12%~+jHA3Z84(h2|bb2;t{RWb(_`h;23{LXH8CxUIJr%w1+b;
zwHurboR;0Lxve^@RbDq(?ubA31|0|SXWHgj!*h4qZp+h-uCA^?nV_Z8c6rgiOXW-O
z5wjFED5bmDn5|QqrbeAUp|no(b&eUBaEtFVuz167a0<t0JmkDdX`dCwo)^KM(gJgI
z6PwXGEHBVH%U|uewdVf7PBqRg*PJ)==EcTo@oY11_RRcR-)=U%gUVW_lG-}AR`)$W
zXAS3Py~9dr;5_Adr&j7hx5}AU=4rd`41VF~O<d!!oWYz@$7<T{woBdpq?b3{e%)-=
z8dgeg)K+ng$EA$DvWL83nJ;wRS}k?<R<a$dwd+<YdiFU#A3KV{&%=bHTLlzx0fS<Z
zi;!m^(`RJt{Or8L{qh=&-{rzHhhk&B18VfSI6mo5H|PluZ07lTKF*~)NM(CWUmW0k
yoBB($1gwY#N~BZmzC3fhIlI%Dt9HBXeq{IxBYGy0OeEEK7yv)`76jPA|B@f^g&A@H

literal 0
HcmV?d00001

diff --git a/test_data/data-multipoint.wkt b/test_data/data-multipoint.wkt
new file mode 100644
index 0000000..d28b0a8
--- /dev/null
+++ b/test_data/data-multipoint.wkt
@@ -0,0 +1,4 @@
+"col","geometry"
+0,"MULTIPOINT ((30 10))"
+1,"MULTIPOINT ((10 40), (40 30), (20 20), (30 10))"
+2,"MULTIPOINT EMPTY"
diff --git a/test_data/data-multipolygon-encoding_native.parquet b/test_data/data-multipolygon-encoding_native.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..00264a4c45067e323bd8543672e45e1a393719eb
GIT binary patch
literal 2411
zcmd5;-)|yS6uvV|>CjkHH{8jP@POFGCbAT0r-7!~aG|9PLYGAg6k|4>VVoV7fsr2t
z+%(z8KI}su{R4dT!9QRh`~!@>_^vU=hkY<6Ci-I3bMHWHO^rVqlRbfZ=ic*u=ic+3
zGlUAMCa{1rc(a5>jBX(mLTH&o+_JdCBiB5$86)h+aujeY*o(bES)6-haRxyWTDxW!
zyi2PzoXGK>Z&Bnq7sNt9LEe>s?9qE2%uAy5oRhxcB8`anZvqFRaC?M#1j7nU2*KLH
zl@flLWs^c^ffMd#78V7xyDXwj_^$dy;Yr>nuAx{42w?sU(DS$$#oHITF~^g9W}cfr
zBk^l)o>z7O_u82cM&B<#{`|#Pzx|2?Sfm7=^Z3rZ&cpBWPq!-P2t~eJ2;wC-=`+zZ
zdbTm_Plb*(7%mze!#29Zzg}<?kZ4=_jF&#)Bj-ZoBM*r5k8gObNaq}GmGCX5Qh}+o
zdPODwOeF!`rBCFtUTLb=0j~Z>y<Y$DTD=6{%B!JQlA$=kwYb(_1I?Z--9WP`)9fc+
z`kasaAV$7lys2h^X9|+N1O1_4A7cSJ?KaQa7b(1aQBlleZ)pc=qX;r9D0C&z<~6>d
z8%|xijbv^zo2N(;1Hr(~|N2V3P_6qy^@Q~Cq9jH{>1(eap+Hpjz<Vt}<sOGmjQ+st
zb;H}Ca3Yq7#lz81*y&sLaew*%e(**%bIpGRsMp;c0ON7^VxKOIuHI`|U6aPLN1dVN
z^g2@$VgqRod@!6k1_0g(mnot$L%e%>dWz(iNVR&ezU}s{{Hb@HTGGrBLhi$t+M9%I
zlB`McgqS2}vb}8fihoQu_b&WphJiI0i^BXWA!7nOm{|~*3E8Csp@eL=Sx=ERT^0sb
z%dSoJWamV))GalqPD<^)Q8}j$)Le0_+nwQ2qo_2J2PaygHgqvX;HQVpMs;eu#c*vc
zIZ0@B<pZ_dGmYsL|CS18H0ugrs*Yys);)_@TB9~-?RUn&Q<};2NVUeMYPULCGSO)j
z;Jl5hqVFF9Hmj+R5(%wPbq@E_1Eb!8dlp;oH4<rheU>846mo}=kc0Lc>+30a*eipG
zgiv*vUNBO9KwL1UM>APF(dxCh-hD*(h+ctu#NUR2^2y)w_4$3eZV`Hvzwsx2c3v-t
z$j)a)VpjugN6ih&-7N!io$(nwenc+-%>c&mEQ>l_y7AckY284>cC+E05BWnMQ2g!@
zjYr{R!p1{7GJV+6eKaVv+y^AhSEXdYF?|ZOf57fvBJHv!bM<rkjUp(EWH$*pqEheE
z*|k670p*%_dpS4WC4Cr?^(XYon8t&JNmFL|GB?Y_Jpsuf=eYm=$Qa(y9p}z*zuy~&
aW-ojwp~T{`cwi@pko!R?FG3agoAo>Rzz6dH

literal 0
HcmV?d00001

diff --git a/test_data/data-multipolygon-encoding_wkb.parquet b/test_data/data-multipolygon-encoding_wkb.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..8eb1edac95ae9887d4b17b8e593e0f939f9f93a8
GIT binary patch
literal 2258
zcmeHJ&2Jh<6n_J2y0~7IsTj0K90+oeRbgr<SOz%}1K435J8`IDFl}WS*x;qCp@2<j
zWXWnH_1Igkm8#x*=)vl#xBd}5RaK-)QBUoyr@onGo!HGKhu-qU%$xT;^P73VF0ZDT
z!U`<2%~S^O0z?5c0z^~SJ>Zac#u>06(-Y_yU_R!H=&Ebc)tL)~VCfZFw9YSzbSm+9
zpF#XNi7+Lc1z#Yndt3b;^MusUa}xTU#OrbObrOdYxO-q8z$t(X0Oh|__SEa1=O-oj
z5Y@y-=7<7bG8aH;4k%G>75wwCsHsrS<Nh=G0MI3>hYRo{`Gj=AL;SQ0ivK{tr0;8a
z<$SZz0Qx0uD47D=312``y9q%}rCItdRW+N*Xyq-ke4}vzw1K)R{h6xl2f74h{{`8*
z!Tf~Y1g>4sZ|I*iKvk-1e=u9CP*pwWh2{~Vi|lW@4ms@!ZLehtez|p=1YN&?RapAk
zLwslv2Pfh-2oBuujLHRP;qr_lpKo5}aO;|1xge8?tRlNZZZI!6^5TrU!LICbS08lz
z+@dhc<8nOqiw(Om*L)y3tL5r**J@`9x?chT4qbU|zq}n`3$xdzcFewQj)znE<o?k@
z->LCmeev(TkvCwz0RLq$U)5pZ^(^SVQR~RG51E23_nt@gF|Y+(U1AI&H5>^i{?~So
zzqwuK$-Axbdad%`+YLQ?C#1&J(3ANfKscd$v19bF$m7_lIUHI2<5)f#OC?jul~^Jg
z8w@S`a5z1{@9+~_n%w>pR!2L+h{v%@y%5afR=;f>cSNq*oucq{?FZxOz(oDWv7M9N
z*xK#)rX4K$>Dk#C=nK4By;sXS4^RKi`+=Y|FLKWB;}CR*^9(O`_$KEaUg}6)5glH#
zPlR&g(!MQCC?m<Gv>TjHL}KJJ33ATMV&XeEipP!G8t(|#%E)Tld(&3BciOa!Rik7)
z+U{<g?39d=QQDfc?B004zLl+~@18a*dt=9E3*&R+M!h;UKazgkX8KX8S<8N5bo(81
zdd0qKV2wsCi$2vs({5S!EN<<sj7x?!=@@pq*G#8+?F!cTu$pb%et`C3)A%fvYF4U)
z2e)%0v)0DFZM7fPQ#pAgUX<q%!123f1aS=)8ygaYl28hKToUIO=bi1>K@#ovCGcWG
zurc30FHR&lBG|>d$d6>6mY8KdDLPQ@r*k8LFXD>vE$);S4ww-M^yuJl`1y%BUTzHr
h%ZI~Ze-fR&!)Pj-TuH8k6A=LXz*GDc*~Nb|{{m?GS)%{|

literal 0
HcmV?d00001

diff --git a/test_data/data-multipolygon.wkt b/test_data/data-multipolygon.wkt
new file mode 100644
index 0000000..79b971f
--- /dev/null
+++ b/test_data/data-multipolygon.wkt
@@ -0,0 +1,5 @@
+"col","geometry"
+0,"MULTIPOLYGON (((30 10, 40 40, 20 40, 10 20, 30 10)))"
+1,"MULTIPOLYGON (((30 20, 45 40, 10 40, 30 20)), ((15 5, 40 10, 10 20, 5 10, 15 5)))"
+2,"MULTIPOLYGON (((40 40, 20 45, 45 30, 40 40)), ((20 35, 10 30, 10 10, 30 5, 45 20, 20 35), (30 20, 20 15, 20 25, 30 20)))"
+3,"MULTIPOLYGON EMPTY"
diff --git a/test_data/data-point-encoding_native.parquet b/test_data/data-point-encoding_native.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..3d650c583b9036ecdb9a98df218a791925579e29
GIT binary patch
literal 1815
zcmcgt&2JJ>5Pu6B#Kr?@+|813AcSsuXc0(h+NL%<=m%S=1scnTng+Jpbt@n1LV>lZ
z{tx~c9{f8z8c#iYHpZhT<IG#2tEGt^?1Y_pJHMHE?>F-{6lf_S@IqLK=E0CzLIUe#
zB;3dfM3@!C1;7J>Pw<5#{}_?{<5-c*Pw@hJH|zn;bDY;BUveRV52lDO5R|-Xrzv<u
zUi`?3pEx-u`~RhG5Klt{4*?YbL&$Uz8)<+*e!RU-!a+9DV-g!_Byrsf9yf_Ed`Osa
zNX+|nt<%zamdy=aTt0i{cV1v~z0?gG{4JYk+9`Q6o9Kj1)Wyc_{e#WVJJZ<ME;iqM
z@&}x`v$ea_6&tsgOk)##mLig`tG>`$b%BRr&3V`jASsVDX(9qMb422XBWn|p4~By4
z|MlX?w=z)qRvIl{@**$$#qa!m0&|eOa6hFt=aIf<b&XCtvKEdcVu@HhvJj4#meHzP
z_7QxCXDtSeFEgT!w+k8Pk%=E)w6@wgHrhJYGCM}QhtcDRqn>SQfIN*981($&;(|!`
zXsNVcUUQ}|edQmZtmqj^X%QZjN0ctnj83;H)oE5|z6=Do@BuovCU~C3z&i+wVz@%-
zfI=QvRs<}io2cj>JehW_yh2Y9%5{z7)}gH?oBf8NEGt=Mm_OY*+si6lCA&LNTg_gz
zvYV<TANL!%!=4kf3;C<PTBT%b2Q2QiksKx(<<x8Cw4-bGZ}_qTd(_G)h$)$kmTK%9
z)KGU;%|^Ru<g?VMB@g?J@?l(UH&2dp%|Yz|&avG!o{VBDmBi`^a>MiH8A-;sR)GE?
z-D1278k+~g@GP#%s50YzP<pk`OGL0gPzU2Tt2heg95Vd`H4bJZY^h=Bo`pk$g@VZt
z$NHPYeuETUNt^7DFia4(#D%lWEKypeaL66Bj$GA)lqvPLvQ9m=kE<S=C-ioJ_25NE
z%52`OE@MI;5-DuftygDSZ&5YP#kys62I0}?5Kg3G@mM^#9wNl~k@q|y1^7$+1<8L`
AK>z>%

literal 0
HcmV?d00001

diff --git a/test_data/data-point-encoding_wkb.parquet b/test_data/data-point-encoding_wkb.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..32a9206c70cc972a8ab1cbd0a6e0850dc5338bc3
GIT binary patch
literal 1375
zcmcgsO>fgc5M9S6N>LA?Vl7+B!LmZsLmJ|MD_T`m*QPY?2Tg<8gd$WWj%!@Sad6zE
zp;Y|?oH+3h_$!DLM~?jo94k)DZsN3{UV5OTc*i^M&8*+El?th9SilO-ZbCvc2&I?M
zIJj}f2+v?K4|p0UaWW&NVnj;CToKLQPz7{KNx&*_9G}Sj<T6;0b(BoY5^o1>oDc=^
z8z=tYa^+m=HgjeAGzcdUENMsx!M=r!SVT#}P~$lCfD@KirtS;bl|>n4FVCS0ae4l)
zrg=1rfgZ;dQMz7UUKtUxG+$Tt>0Sm;kCmldC-9vp95Q)d65?+z_lwVc<}<fG<R+Q&
zl8&Tg&wl6l4J?4}M+s&o!qad!IUI?CoRRbY)$`n)$rdkoQOKpl@BBlA<h;a#$I{1B
zweN&IFX*U?imDZ~LQ%~tYB%)!MmRix5A4`yQ2atDI;!@d<5Zp0>Bi~ULCfp7bXIM{
zpxhWgh=yGU>TlFa;B_MP?b+EGlBP+ux?Niwy$R`pe@vr93_{2bTr@Tac}fhIydcCS
zrpsyu5U$|^T3nsrn~Vc{2*zT#MaX~-hRlXQCgeF?$QE2iXTSE09MWyM=e7L3p<QbC
zo1QgqnbyhX(dzM*Y4t2~ePH|T=yiQvua{o-o8`S|#94>_g=oKC9XhWV@2FWi(V8{=
zgLM?R&hRgM-hwmsYdUbM-KKAQ+aB@uijiq~1K09f?Pf`9x5{wFld5j7?E-Ezt+vr`
z)byg&))o%yMf#*8DXUBm(_5^7e^l<y4pYuVPOKutq&;g_^XZnF1ir9-Sa#`9@ViX?
zia~k<6qAvR*zZww>>p?u(D`5{^be@d<A8~AAW7*q!uQ8cG-r3abB!<z2FmyYR5ZO%
TEEMH@1|j&NuRKB(_-p+GpsW(6

literal 0
HcmV?d00001

diff --git a/test_data/data-point.wkt b/test_data/data-point.wkt
new file mode 100644
index 0000000..6f8b211
--- /dev/null
+++ b/test_data/data-point.wkt
@@ -0,0 +1,3 @@
+"col","geometry"
+0,"POINT (30 10)"
+1,"POINT EMPTY"
diff --git a/test_data/data-polygon-encoding_native.parquet b/test_data/data-polygon-encoding_native.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..a5eea03772c7de1d2f0f5bc3e965aa3d7ec6c731
GIT binary patch
literal 2190
zcmc&$-EJC36rNp{_1cOnE6S`E2}KBzm8BB05U*t-H(~(WtP2SyDPXE9G3)IH%3t|0
zY$8ihE~-30udLj(y~z_qRUe@b(2HL7x)0GaXTgpoOKns|Ioh4sne&}*=6rLOrOYZx
zDp7;Ri&UoMIw2Y%tHQQGd=uwACG-Z>;(%F$TMlc=!lNm32vN!U6|>-7iE*$jh=CZ<
z{}RGf3YkP+3uytn-=U%^tA7e`>1AE{5aLiAc21~Bpjm^C5Eu{WgChM5Gua{}C`k7+
z!6k|0R~51WcWhaaUb0mswwnPwSRw<_DY_Y_H_uZ;$IIQ!BDGjX71#YVFU<lj^fU-<
z@a@lMD-t|}1ELTJzLx*^MG^FO!7yF%GkvDGPQUAnho01FjmAq($LTt~@uh(<11WB*
zzl!RasQ)18Zvr63f8NTKVmzgVBE5!U7$`>ULJTg%?y|B@?q~c(*hf&JAjQScMg8~X
z4^zTJCEkkakD`92=<k;PuM(j*Ceh@P{mkheQ3>kgmWahiOi?>O8Y)t`xD8^Ml7=S(
zH3ead!}>Qj(5~DfnlPL99#Q2`IJEsyEB;d#*{42Q^*33SbwzzG-+-4ft_9#d(_RTL
zq9@L9)av)5_af0`BAGCv@kn$qY;})@-V^xMnRL;$cq_TrpB+KOi|F}0A2>a`f86T1
zoNLhUcrMTaet~&1_6834zK)jX&TmdnPl<MoRVoi__xz&M{tn*Z=D9h>*a6(!zGQ5J
zWnH$*n9Fi5&SkTg{8K);bM7x<D2&1I5&afpQwDL+BMJ13<$1$OaA$js+7@%Ut1xOE
zcdMSA>YOxN)~1!SX2tf-q?EHpR&IZ4cRS<5`o39DJveC=s$(Cs5Ao@7qh9fxM~G`T
zQ?q2VW<Iyteb@0W`8O?Cqfs*fQyDb7cI#n_wd}q0$f+IkHTgOheATq!)UzGVQ}gx@
zfY)rK(i5wbOn+N9V8fWo0nZik18akbThQ;6Ks3g<ny3%7Tn`Wv41*e}>PfR!HSFGZ
ze2(?Y5PJ9VJf8XS*@eCUz<3lDsX>cL<1W`PUxJ`J7>{t{kn_Refgv7uQR73O57*~9
zAMbVAjXLMR{lmF&eEyL0;p7b4uXyX{hhuznIJ7hbB+u94MF2R=ufWe6?BKniJQjtY
zzh565!J}AqgRw&{)gJGCyWoS6EBJPDKHq!#U?A2Lob8nJ!NBDyW4`FdOxPDhiwus2
p-%p(JZF?}deKZ{Qr;+(98A+N6BVmNL!|=Vti~Iy1(K7rK`3IsUzoY;F

literal 0
HcmV?d00001

diff --git a/test_data/data-polygon-encoding_wkb.parquet b/test_data/data-polygon-encoding_wkb.parquet
new file mode 100644
index 0000000000000000000000000000000000000000..4ec68a6680afe4c69bcaa4f92f24d2e0d610e70a
GIT binary patch
literal 1857
zcmd^A&2HL25FUdeCJGX+vdEUGgb+ojA`~b@Nfbpa<j?tO2vq`28%01I>WIGr387Ic
zPJMx1d&#A$sy;#=pwG}#m2)4Wvx`kil}isjR2_MDyz_lKvokxgb2J|(B8id67I6}|
z2cQ7Bg+&2|Vly=XGEI~Say&SkK1H$+MKUSih439#L??rc>=bPFAcSrN9}(3QI6bOj
zuQY4K=5>0%3HXF^A?Z&Xs}Vdo5F6m+K?Q*84<h*?Sh>8jO&_9JunP_kY6r-IxRiEH
zLeMS2@&cYbx!?ggEGudnu7wvu32f04hy&`T?s;*r>hj2v2vL`Z^xa{(bUzKE>l4QD
z$yahvmSja<Cej?b<WX>9@_VgjA$%<ec0|}29gM<cK+{}9aHt2X;~JtSzHDlf2T7B6
zyI%xLObnLBIipyn-SRQgOKHW$k7zm?yRxs6k4%|YCRO!jUGKL0{N8RRyq_JRkB-oL
zyYJsi^R5O`7gL}(JC&Duy+lOps0TLQra-3fZInHDovKfb{MBv_{Xcd0esOu7A<6r8
zVj4h=D0b{J<;`W_TyJ-bW+SlR55%MKXe<!%2U=~TUTXJ`@Y``#=aAKs3>{iKxZ^S~
zva^NWs5Fm_MwR(0Y-SkO!jHQBmX7x4fn2lJuVTEztE($erf5FDQ&<=_gYw<}h)IHG
zD5bkNm~2v-q?syRrL;=3Rc^~5;TOEej4LC2i*s;|!dV>GDD5%FsJRo=lrFQ0ZsW)_
z4hnPhghA^a<G8-xuOw>cWkZ|Qvf9Pg>B`x5R_kcljb5c*>pm-Pq>70r=jHYN?vS&A
z`{%m{#e83X%6X^d#6`SZNWIohn^nC(hR<pk<Dig2PQF#HSBxEl8vC(sRx^54t$ti9
zC*rl^b&T;ipQ^0wA)YB~wG76qHOh&kQ7p#iPgrj3yem8(o^LFN^<(Mo?()oelA}^l
znnj4N-TG%sRwS$o-;dKCbBgs|XZ5l`_JWiwCuOX953OeXFqa<lkHy6OJyvH^(4!@w
q_*<p+t24bjQ)#tkO6_*D=Qmr>A5TSN(U=<X0pJ4<@td2&pXX2H*EUK3

literal 0
HcmV?d00001

diff --git a/test_data/data-polygon.wkt b/test_data/data-polygon.wkt
new file mode 100644
index 0000000..c528373
--- /dev/null
+++ b/test_data/data-polygon.wkt
@@ -0,0 +1,4 @@
+"col","geometry"
+0,"POLYGON ((30 10, 40 40, 20 40, 10 20, 30 10))"
+1,"POLYGON ((35 10, 45 45, 15 40, 10 20, 35 10), (20 30, 35 35, 30 20, 20 30))"
+2,"POLYGON EMPTY"
diff --git a/test_data/generate_test_data.py b/test_data/generate_test_data.py
new file mode 100644
index 0000000..3324806
--- /dev/null
+++ b/test_data/generate_test_data.py
@@ -0,0 +1,195 @@
+"""
+Generates example data using pyarrow by running `python generate_test_data.py`.
+
+You can print the metadata with:
+
+.. code-block:: python
+
+   >>> import json, pprint, pyarrow.parquet as pq
+   >>> pprint.pprint(json.loads(pq.read_schema("example.parquet").metadata[b"geo"]))
+"""
+import json
+import pathlib
+import copy
+
+import pyarrow as pa
+import pyarrow.parquet as pq
+from pyarrow.csv import write_csv
+
+from shapely import from_wkt, to_wkb
+
+
+HERE = pathlib.Path(__file__).parent
+
+
+metadata_template = {
+    "version": "1.1.0",
+    "primary_column": "geometry",
+    "columns": {
+        "geometry": {
+            "encoding": "WKB",
+            "geometry_type": [],
+        },
+    },
+}
+
+
+## Various geometry types with WKB and native (GeoArrow-based) encodings
+
+def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
+
+    table = pa.table({"col": range(len(geometries_wkt)), "geometry": geometries_wkt})
+    write_csv(table, HERE / f"data-{geometry_type.lower()}.wkt")
+
+    # WKB encoding
+    table = pa.table(
+        {"col": range(len(geometries_wkt)), "geometry": to_wkb(from_wkt(geometries_wkt))}
+    )
+    metadata = copy.deepcopy(metadata_template)
+    metadata["columns"]["geometry"]["geometry_type"] = [geometry_type]
+    table = table.replace_schema_metadata({"geo": json.dumps(metadata)})
+    pq.write_table(table, HERE / f"data-{geometry_type.lower()}-encoding_wkb.parquet")
+
+    # native (geoarrow) encoding
+    table = pa.table(
+        {"col": range(len(geometries_wkt)), "geometry": geometries_geoarrow}
+    )
+    metadata["columns"]["geometry"]["encoding"] = geometry_type.lower()
+    table = table.replace_schema_metadata({"geo": json.dumps(metadata)})
+    pq.write_table(table, HERE / f"data-{geometry_type.lower()}-encoding_native.parquet")
+
+
+# point
+
+geometries_wkt = [
+    "POINT (30 10)",
+    "POINT EMPTY",
+]
+
+point_type = pa.struct(
+    [
+        pa.field("x", pa.float64(), nullable=False),
+        pa.field("y", pa.float64(), nullable=False)
+    ]
+)
+geometries = pa.array([(30, 10), (float("nan"), float("nan"))], type=point_type)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="Point"
+)
+
+# linestring
+
+geometries_wkt = [
+    "LINESTRING (30 10, 10 30, 40 40)",
+    "LINESTRING EMPTY",
+]
+
+linestring_type = pa.list_(pa.field("vertices", point_type, nullable=False))
+geometries = pa.array(
+    [[(30, 10), (10, 30), (40, 40)], []], type=linestring_type)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="LineString"
+)
+
+# polygon
+
+geometries_wkt = [
+    "POLYGON ((30 10, 40 40, 20 40, 10 20, 30 10))",
+    "POLYGON ((35 10, 45 45, 15 40, 10 20, 35 10), (20 30, 35 35, 30 20, 20 30))",
+    "POLYGON EMPTY",
+]
+
+polygon_type = pa.list_(
+    pa.field("rings", pa.list_(
+        pa.field("vertices", point_type, nullable=False)
+    ), nullable=False)
+)
+geometries = pa.array(
+    [
+        [[(30.0, 10.0), (40.0, 40.0), (20.0, 40.0), (10.0, 20.0), (30.0, 10.0)]],
+        [[(35.0, 10.0), (45.0, 45.0), (15.0, 40.0), (10.0, 20.0), (35.0, 10.0)],
+         [(20.0, 30.0), (35.0, 35.0), (30.0, 20.0), (20.0, 30.0)]],
+        [],
+    ],
+    type=polygon_type
+)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="Polygon"
+)
+
+# multipoint
+
+geometries_wkt = [
+    "MULTIPOINT ((30 10))",
+    "MULTIPOINT ((10 40), (40 30), (20 20), (30 10))",
+    "MULTIPOINT EMPTY",
+]
+
+multipoint_type = pa.list_(pa.field("points", point_type, nullable=False))
+geometries = pa.array(
+    [
+        [(30.0, 10.0)],
+        [(10.0, 40.0), (40.0, 30.0), (20.0, 20.0), (30.0, 10.0)],
+        []
+    ],
+    type=multipoint_type
+)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="MultiPoint"
+)
+
+# multilinestring
+
+geometries_wkt = [
+    "MULTILINESTRING ((30 10, 10 30, 40 40))",
+    "MULTILINESTRING ((10 10, 20 20, 10 40), (40 40, 30 30, 40 20, 30 10))",
+    "MULTILINESTRING EMPTY",
+]
+
+multilinestring_type = pa.list_(
+    pa.field("linestrings", linestring_type, nullable=False)
+)
+geometries = pa.array(
+    [
+        [[(30.0, 10.0), (10.0, 30.0), (40.0, 40.0)]],
+        [[(10.0, 10.0), (20.0, 20.0), (10.0, 40.0)],
+         [(40.0, 40.0), (30.0, 30.0), (40.0, 20.0), (30.0, 10.0)]],
+        [],
+    ],
+    type=multilinestring_type
+)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="MultiLineString"
+)
+
+# multipolygon
+
+geometries_wkt = [
+    "MULTIPOLYGON (((30 10, 40 40, 20 40, 10 20, 30 10)))",
+    "MULTIPOLYGON (((30 20, 45 40, 10 40, 30 20)), ((15 5, 40 10, 10 20, 5 10, 15 5)))",
+    "MULTIPOLYGON (((40 40, 20 45, 45 30, 40 40)), ((20 35, 10 30, 10 10, 30 5, 45 20, 20 35), (30 20, 20 15, 20 25, 30 20)))",
+    "MULTIPOLYGON EMPTY",
+]
+
+multipolygon_type = pa.list_(pa.field("polygons", polygon_type, nullable=False))
+geometries = pa.array(
+    [
+        [[[(30.0, 10.0), (40.0, 40.0), (20.0, 40.0), (10.0, 20.0), (30.0, 10.0)]]],
+        [[[(30.0, 20.0), (45.0, 40.0), (10.0, 40.0), (30.0, 20.0)]],
+         [[(15.0, 5.0), (40.0, 10.0), (10.0, 20.0), (5.0, 10.0), (15.0, 5.0)]]],
+        [[[(40.0, 40.0), (20.0, 45.0), (45.0, 30.0), (40.0, 40.0)]],
+         [[(20.0, 35.0), (10.0, 30.0), (10.0, 10.0), (30.0, 5.0), (45.0, 20.0), (20.0, 35.0)],
+          [(30.0, 20.0), (20.0, 15.0), (20.0, 25.0), (30.0, 20.0)]]],
+        [],
+    ],
+    type=multipolygon_type
+)
+
+write_encoding_files(
+    geometries_wkt, geometries, geometry_type="MultiPolygon"
+)

From 7e9eed264e68fa417863ce65f7bf5f7be8b302de Mon Sep 17 00:00:00 2001
From: Joris Van den Bossche <jorisvandenbossche@gmail.com>
Date: Thu, 2 May 2024 16:15:19 +0200
Subject: [PATCH 2/8] Update test_data/generate_test_data.py

Co-authored-by: Even Rouault <even.rouault@spatialys.com>
---
 test_data/generate_test_data.py | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/test_data/generate_test_data.py b/test_data/generate_test_data.py
index 3324806..f5915c0 100644
--- a/test_data/generate_test_data.py
+++ b/test_data/generate_test_data.py
@@ -28,7 +28,7 @@
     "columns": {
         "geometry": {
             "encoding": "WKB",
-            "geometry_type": [],
+            "geometry_types": [],
         },
     },
 }

From c1f11ebdf29fdf94abcf61c40af7f677967dc8b9 Mon Sep 17 00:00:00 2001
From: Joris Van den Bossche <jorisvandenbossche@gmail.com>
Date: Thu, 2 May 2024 16:18:35 +0200
Subject: [PATCH 3/8] fix geometry type

---
 .../data-linestring-encoding_native.parquet   | Bin 2017 -> 2018 bytes
 .../data-linestring-encoding_wkb.parquet      | Bin 1468 -> 1469 bytes
 ...ta-multilinestring-encoding_native.parquet | Bin 2211 -> 2212 bytes
 .../data-multilinestring-encoding_wkb.parquet | Bin 1802 -> 1803 bytes
 .../data-multipoint-encoding_native.parquet   | Bin 2022 -> 2023 bytes
 .../data-multipoint-encoding_wkb.parquet      | Bin 1616 -> 1617 bytes
 .../data-multipolygon-encoding_native.parquet | Bin 2411 -> 2412 bytes
 .../data-multipolygon-encoding_wkb.parquet    | Bin 2258 -> 2271 bytes
 test_data/data-point-encoding_native.parquet  | Bin 1815 -> 1816 bytes
 test_data/data-point-encoding_wkb.parquet     | Bin 1375 -> 1376 bytes
 .../data-polygon-encoding_native.parquet      | Bin 2190 -> 2191 bytes
 test_data/data-polygon-encoding_wkb.parquet   | Bin 1857 -> 1858 bytes
 test_data/generate_test_data.py               |   2 +-
 13 files changed, 1 insertion(+), 1 deletion(-)

diff --git a/test_data/data-linestring-encoding_native.parquet b/test_data/data-linestring-encoding_native.parquet
index d4853894ddbcdb21afea8c2d2d3f14f667f9a885..ee736c1fb71d8b7a4ae8e00d9e3714e913121750 100644
GIT binary patch
delta 68
zcmaFJ|A>Es9}8pWWPg@Y#^T9GSe7#ePR?PSCzD+1Y3AuxoEhStQyLZKXONr=q$|Br
X(jp8kH><I2X5?^YWnc(!3^D`&JnR+_

delta 66
zcmaFF|B!!!9}8p0WPg^@$wyh1GdfPrWt}JEnU$X&Ug_nV7;c&qY8+Gvr1Ly;3@y_l
V3>-JBvu$SNaAIX(2yhHC1ON`H6$k(T

diff --git a/test_data/data-linestring-encoding_wkb.parquet b/test_data/data-linestring-encoding_wkb.parquet
index 853ac9ee231e6c46f9f9cc3b9df90acd7dc529d4..9190fc0da040f4ec3fcdb8c063cc8dd92f086b43 100644
GIT binary patch
delta 68
zcmdnPy_b7~Ff(J*WD({(#^T9M%*z?mCu^}xkqxWz%*sy>uk`Xw3^&aQH4dr-(s`aa
XhL&j&29s~I+~hdM%)k)f7-R?l{qPn7

delta 66
zcmdnXy@z{)Ff(K0WD(}P$<55m83QM4vrLf<&GgT7O7hIEC`if;3rX?t0@BHu5fO%#
VX@QQD@37qDILgex5a1YO2ms>C73Tl|

diff --git a/test_data/data-multilinestring-encoding_native.parquet b/test_data/data-multilinestring-encoding_native.parquet
index fb4780abdb7b6b28bb2c3d50fe8daadac94060fc..04e2e7c0f6de2b1dee9bcb3a2e47dca08c8559e1 100644
GIT binary patch
delta 76
zcmZ21xI}QnE*8eglXtUJF&0lYWnIHqKKT*r5|ywj&#e6P@JcVgl<*3J#L$X@q};HO
f6b~;TotzmFVQ85axOo-ZM@EiItPBhRjzNY1N#Pqb

delta 74
zcmZ1?xL9z*E*8c~lXtUJO*Ug)!<at#G3yeQ&`keKrzFp8L&GGGpn?$hoYJT;KZE34
dAYJK|k``fT>FBt5HQPr<j*F}e3;~Woh5-HK8O8tr

diff --git a/test_data/data-multilinestring-encoding_wkb.parquet b/test_data/data-multilinestring-encoding_wkb.parquet
index 1357eb2e416c338e6aae2e6e0eab8165ae496cd2..b3e7b26f1f88cf989d57c08facecd0c6fe5c1e74 100644
GIT binary patch
delta 75
zcmeC;>*m`Kz{1!*Igllfv3T-+mNks!lXF<7sDxE{X62`cS9<xSgjX0OhE^0L<%Wf%
dcz6Nn<jjZ&L(8<l$trC3IJ8(87y=xF3<20_7_a~U

delta 73
zcmeC?>*CuGz{1!zIgll9@&T4LjOmkeS*NIkX8LD3C3$8W8YX!J6@<9wltzX586@Wd
b=}NDZv<O2>N5{#kZ1*@cSr`}s9D@u2kT4fQ

diff --git a/test_data/data-multipoint-encoding_native.parquet b/test_data/data-multipoint-encoding_native.parquet
index ef9791ea9faff1a1bbd88f4a9aba50174a96337f..78bd1329a1cebc68b456cccaa614804a2f7fb8ef 100644
GIT binary patch
delta 68
zcmaFH|D1nAGz(+r<QSGx#^T8rS(Y;fPOfB~CzD+1Y3AuxoEZ`kR-EFV6PRR_Q<~!J
Ym68@=Xt~*#Z8al@H7f%{fMbv$07kkN761SM

delta 66
zcmaFP|BQb_Gz(+L<QSIH$(LA`GdfPLVx1@BnU$X&Ug_nR5?*1D80J)#7;b9dnPX^~
W7GdDH*@SI1BZn0$14DpgkRbpYkQFQd

diff --git a/test_data/data-multipoint-encoding_wkb.parquet b/test_data/data-multipoint-encoding_wkb.parquet
index 9fd760557604428b15690059d31ea5e8c7903f1f..6583b3e2531a1722560ea61740aa5e0ebd663c37 100644
GIT binary patch
delta 68
zcmcb>bCGAma%RS+$t#%i7>g&1u`FjypS+J{ifmYwXI6fCc%_$LN_d4qVwh7|Vz{Y+
YXO5v|T7<#meAb&Bub3Ga0vv-30Z()n9{>OV

delta 66
zcmcb}bAe~Wa%RTH$t#%iCX2HyXAGRYpJj?{Xr_OrQ<7)4p<$9oP(i4>WkFJ2U}i*w
Wp=Da2<KzO?n;b8h85jZ_gA4%~+7?y-

diff --git a/test_data/data-multipolygon-encoding_native.parquet b/test_data/data-multipolygon-encoding_native.parquet
index 00264a4c45067e323bd8543672e45e1a393719eb..060beea13a3db5328d34911d52ce025a14000fa1 100644
GIT binary patch
delta 72
zcmaDY^hRhy2rFan<WSaf#^T9mSXVLzPA+3xpb%E&nU$X&Ug_nR5?*1D80J)#<dKt?
cWMq;V5n*VV7P#4rU6+Z&l8u2Oz%j@W0I?t!EC2ui

delta 70
zcmaDO^jc^`2rFaH<WSb~$!A$tGCEE!XIr2Un(3eEl;oLhXqe;?R1oTJS)3Z4QkInI
am68@=XzA#<*_>UMiNk`8fg!*#$PfT^bQUcD

diff --git a/test_data/data-multipolygon-encoding_wkb.parquet b/test_data/data-multipolygon-encoding_wkb.parquet
index 8eb1edac95ae9887d4b17b8e593e0f939f9f93a8..65a3cec73577ae336568c10043aaf909b93d52cd 100644
GIT binary patch
delta 275
zcmca4cwcaXIvZolWDT}F#^TA-*wi@RFzbVWQTpVgZ1#-llUdm<nKB$FJFriw53BOb
z%1;lk^zusyuP{gqb1F;n$jM7GGRcgHFtkhybOZrs5D^R_96<sufgsa>3@0ZikeFjS
zh=4FbEJsfe5oHusmJ}Xlkertd;)3K|A^L%aI0b?Xb$1MROm=iZ$buC*26}?6^#JoA
rE&wyjfuy@%SrXWSa*zQ)C1ATj(m*5Oz;^OH4jqogEDQ_*jzNY1a@9@N

delta 261
zcmcaFcu8=BIvZp2WDT~w$<x`?I3F<UgMg9a<YR31jDeHc*e#jT%O^XsPpJ>h^v`rk
z^2|0gO!5dS2z9qCP7O~fOUm?0NsBPFbaeD|baZqA5spqE!p{Y0js=k6>go#OJA(*E
zFcZvjOb1d<c~Rk(B~cMUm8pSXQJ{PvL_g3FH?X0Rj!xx{zK-4y-4JCUJ2Jo;l0hWc
sLNM3S1w=&}Sr$72EdX+W4oZgF4Uz#G2?Uc@bLenPV_{$ja11g80078Ik^lez

diff --git a/test_data/data-point-encoding_native.parquet b/test_data/data-point-encoding_native.parquet
index 3d650c583b9036ecdb9a98df218a791925579e29..774dbcced094ab0d1be1ba7526fbef64cd8199fd 100644
GIT binary patch
delta 62
zcmbQvH-m4(V`j$s$xoQ`8H*<evn*mPpZtwwx>$s7rhleWl4p)nSz@@UfoG1PWm<&6
S=EbZzj2x?37#IQ^gA4)2Fcgsh

delta 60
zcmbQiH=S?8V`j#>$xoQ`Cx@^sVoaa>on^XMu&0@)TXANnyJbOAUSMWKgrQ|xpyTEx
QtT~Jvt5_Hq0vv-30ivB0D*ylh

diff --git a/test_data/data-point-encoding_wkb.parquet b/test_data/data-point-encoding_wkb.parquet
index 32a9206c70cc972a8ab1cbd0a6e0850dc5338bc3..e9c5f7402cc7484d35afc98e463f3f15ee41ec59 100644
GIT binary patch
delta 61
zcmcc5^?+-`9wx@x$$Ob{8H*=tF)v~aoP3sfida~cXI6fCc%@fhl2J}+inCWrT7;qH
R<Z_k=94nX^7y=xF3<1DY6(0Zq

delta 59
zcmaFBb)Rd)9wx?`$$Ob{Cu=h=VsxB*j(Lh$Xr_OrQ<7(nQ(0oTsexyXp=DZxf#c)~
PmIoZmnHd-Y9D@u2ioFxq

diff --git a/test_data/data-polygon-encoding_native.parquet b/test_data/data-polygon-encoding_native.parquet
index a5eea03772c7de1d2f0f5bc3e965aa3d7ec6c731..5671f55777273ccf465eb39ee387fb75aa743bfd 100644
GIT binary patch
delta 64
zcmeAZ>=)c{m4&f+@->zs#^T8ytV<ZnC%<K#C6Qd|Y3AuxoEhqFS)3Z4QkInIm68@=
UXt{Yl+jB;aMXU@A0gge20Q=1r8vp<R

delta 62
zcmeAd>=WE@m4&fs@->#C$)2oB7}F=eW1S`8nU$X&Ug;H>WK>}qWn@|EnPX^~7GdDH
Sc>&vVMvjH73=9E|L52X%qZJSU

diff --git a/test_data/data-polygon-encoding_wkb.parquet b/test_data/data-polygon-encoding_wkb.parquet
index 4ec68a6680afe4c69bcaa4f92f24d2e0d610e70a..06f065d1dcfc85d494a3bd3d37412511f5f2b0e0 100644
GIT binary patch
delta 64
zcmX@ecZhGp92Um<$#Yq98H*=#u`XdOpS+%Riey-oXI6fCc%@fhl2L_el#ykrXO5v|
UT7<#m7`D3{+AItV0gge20NAV)pa1{>

delta 62
zcmX@acaU$x92Um9$#Yq9Cv&qdVN9RAfpv;xXr_OrQ<7(nQ(2NnPF|9cNoGWZp=Da2
S<K$SjyBu093=9E|L52Xb>J%RU

diff --git a/test_data/generate_test_data.py b/test_data/generate_test_data.py
index f5915c0..a1945d1 100644
--- a/test_data/generate_test_data.py
+++ b/test_data/generate_test_data.py
@@ -46,7 +46,7 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
         {"col": range(len(geometries_wkt)), "geometry": to_wkb(from_wkt(geometries_wkt))}
     )
     metadata = copy.deepcopy(metadata_template)
-    metadata["columns"]["geometry"]["geometry_type"] = [geometry_type]
+    metadata["columns"]["geometry"]["geometry_types"] = [geometry_type]
     table = table.replace_schema_metadata({"geo": json.dumps(metadata)})
     pq.write_table(table, HERE / f"data-{geometry_type.lower()}-encoding_wkb.parquet")
 

From 51354939942074d3ef7ba22d674596f54209d7bd Mon Sep 17 00:00:00 2001
From: Joris Van den Bossche <jorisvandenbossche@gmail.com>
Date: Thu, 2 May 2024 16:20:49 +0200
Subject: [PATCH 4/8] add .csv extension

---
 test_data/{data-linestring.wkt => data-linestring.wkt.csv}      | 0
 .../{data-multilinestring.wkt => data-multilinestring.wkt.csv}  | 0
 test_data/{data-multipoint.wkt => data-multipoint.wkt.csv}      | 0
 test_data/{data-multipolygon.wkt => data-multipolygon.wkt.csv}  | 0
 test_data/{data-point.wkt => data-point.wkt.csv}                | 0
 test_data/{data-polygon.wkt => data-polygon.wkt.csv}            | 0
 test_data/generate_test_data.py                                 | 2 +-
 7 files changed, 1 insertion(+), 1 deletion(-)
 rename test_data/{data-linestring.wkt => data-linestring.wkt.csv} (100%)
 rename test_data/{data-multilinestring.wkt => data-multilinestring.wkt.csv} (100%)
 rename test_data/{data-multipoint.wkt => data-multipoint.wkt.csv} (100%)
 rename test_data/{data-multipolygon.wkt => data-multipolygon.wkt.csv} (100%)
 rename test_data/{data-point.wkt => data-point.wkt.csv} (100%)
 rename test_data/{data-polygon.wkt => data-polygon.wkt.csv} (100%)

diff --git a/test_data/data-linestring.wkt b/test_data/data-linestring.wkt.csv
similarity index 100%
rename from test_data/data-linestring.wkt
rename to test_data/data-linestring.wkt.csv
diff --git a/test_data/data-multilinestring.wkt b/test_data/data-multilinestring.wkt.csv
similarity index 100%
rename from test_data/data-multilinestring.wkt
rename to test_data/data-multilinestring.wkt.csv
diff --git a/test_data/data-multipoint.wkt b/test_data/data-multipoint.wkt.csv
similarity index 100%
rename from test_data/data-multipoint.wkt
rename to test_data/data-multipoint.wkt.csv
diff --git a/test_data/data-multipolygon.wkt b/test_data/data-multipolygon.wkt.csv
similarity index 100%
rename from test_data/data-multipolygon.wkt
rename to test_data/data-multipolygon.wkt.csv
diff --git a/test_data/data-point.wkt b/test_data/data-point.wkt.csv
similarity index 100%
rename from test_data/data-point.wkt
rename to test_data/data-point.wkt.csv
diff --git a/test_data/data-polygon.wkt b/test_data/data-polygon.wkt.csv
similarity index 100%
rename from test_data/data-polygon.wkt
rename to test_data/data-polygon.wkt.csv
diff --git a/test_data/generate_test_data.py b/test_data/generate_test_data.py
index a1945d1..8fe7377 100644
--- a/test_data/generate_test_data.py
+++ b/test_data/generate_test_data.py
@@ -39,7 +39,7 @@
 def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
 
     table = pa.table({"col": range(len(geometries_wkt)), "geometry": geometries_wkt})
-    write_csv(table, HERE / f"data-{geometry_type.lower()}.wkt")
+    write_csv(table, HERE / f"data-{geometry_type.lower()}.wkt.csv")
 
     # WKB encoding
     table = pa.table(

From ffb547d31ba990fe7f8ebd2ea126ae2a502beee8 Mon Sep 17 00:00:00 2001
From: Joris Van den Bossche <jorisvandenbossche@gmail.com>
Date: Thu, 2 May 2024 16:27:21 +0200
Subject: [PATCH 5/8] add null values

---
 .../data-linestring-encoding_native.parquet   | Bin 2018 -> 2023 bytes
 .../data-linestring-encoding_wkb.parquet      | Bin 1469 -> 1474 bytes
 test_data/data-linestring.wkt.csv             |   1 +
 ...ta-multilinestring-encoding_native.parquet | Bin 2212 -> 2216 bytes
 .../data-multilinestring-encoding_wkb.parquet | Bin 1803 -> 1807 bytes
 test_data/data-multilinestring.wkt.csv        |   1 +
 .../data-multipoint-encoding_native.parquet   | Bin 2023 -> 2027 bytes
 .../data-multipoint-encoding_wkb.parquet      | Bin 1617 -> 1621 bytes
 test_data/data-multipoint.wkt.csv             |   1 +
 .../data-multipolygon-encoding_native.parquet | Bin 2412 -> 2417 bytes
 .../data-multipolygon-encoding_wkb.parquet    | Bin 2271 -> 2276 bytes
 test_data/data-multipolygon.wkt.csv           |   1 +
 test_data/data-point-encoding_native.parquet  | Bin 1816 -> 1835 bytes
 test_data/data-point-encoding_wkb.parquet     | Bin 1376 -> 1398 bytes
 test_data/data-point.wkt.csv                  |   2 +
 .../data-polygon-encoding_native.parquet      | Bin 2191 -> 2195 bytes
 test_data/data-polygon-encoding_wkb.parquet   | Bin 1858 -> 1862 bytes
 test_data/data-polygon.wkt.csv                |   1 +
 test_data/generate_test_data.py               |  45 +++++++++++-------
 19 files changed, 35 insertions(+), 17 deletions(-)

diff --git a/test_data/data-linestring-encoding_native.parquet b/test_data/data-linestring-encoding_native.parquet
index ee736c1fb71d8b7a4ae8e00d9e3714e913121750..32653ef94bb29fab1ec39a36a082423323fa53cf 100644
GIT binary patch
delta 405
zcmaFF|D4}Cz%j^Bltt7)RL4h@O_V{1fkA?SftS&SkpTjjptLB1sEDY94p5B%5VOfh
za4=2m7j=P2OK?EdfhaB&kT4q~6SE3~+9O68Q5H#41~vgn3FhSd98o4QHnCfbK%^F+
z#sM{lk!i9%v&!TRj3N{7N-%OwzRM`X$TGQrNpvzllK=<ApcdB24veZiTw*U6fk^EM
zv)Y`=g^c20b(#}T3r#-4h)e4w4xm;=;Yy(0FzuVz)m{O8F}aXQm`My|Rig%*q>Q8u
h3GP*8!sjG)W;J50Rp%62#VW?8#v!(9vlGh#W&od#OZ5N%

delta 426
zcmaFP|A^l=z%j^BltolQRK*8KF$gg*2-q+{0V9+aWe^n-71IIA3y89bvdKtrz+@yi
zU~0t}G&mS128y_FDlh@XSQr_ZnbdAD%80T^nli8nNJ=m#=jVtriLr>CU=+K+sOF)@
zfybOx0gN0If6Fj3P5#fQ%E7?Itiqu7i)HdTW<|jYpdlP$FBpMH?H04zB<9JCjACGg
zqLX!4geIO7nS7E_0!ceNMEi-!XPMP_U^-W@t38;+D2~uKc@9&Q9gC!lqzwhGR$(U6
Zl}u{*og>UKgH?=0jYDj~W_Oka%m67yPO1O^

diff --git a/test_data/data-linestring-encoding_wkb.parquet b/test_data/data-linestring-encoding_wkb.parquet
index 9190fc0da040f4ec3fcdb8c063cc8dd92f086b43..276e4184eda51af7ef65e67abb4c7d54ba3e88a1 100644
GIT binary patch
delta 296
zcmdnXeTdsSz%j^Bltt7)RL4h@O_V{1fkA?SftS&SkpTjjptLB1sEDY94p5B%5VOfh
za4=2m7j=P2OK?EdfhaB&kT4q~6SE3~+9O68Q5H#41~vgn3FhSd98o4QHnCfbK%^F+
z#sM{lk!f-Slgh-CZd@?)&6p-jGWxSKGcz(XsXdvT&nPayCiaU->;#kA3>LLH%ramF
n?vw8_+G?;#%1G*vV38^#UIT?V#jdc3v8i#0E!xb^yp0I}=u|N5

delta 362
zcmX@ay_efJz%j^BltolQRK*8KF$gg*2-q+{0V9+aWe^n-71IIA3y89bvdKtrz+@yi
zU~0t}G&mS128y_FDlh@XSQr_ZnbdAD%80T^nli8nNJ=m#=jVtriLr>CU=+K+sOF)@
zf!mzPu}mrx&$)5IEHq=7EXU}t2Ql;(tBfcQ)KHG})coAklA=nm$-kJyPB5wUv8YXA
rmH}&VpZti?){aF|MpB0ox2S56Xr~aT*dZ1%7BvpB8Jl^Sw=n?#e#|vt

diff --git a/test_data/data-linestring.wkt.csv b/test_data/data-linestring.wkt.csv
index 26f6fa4..a3a320d 100644
--- a/test_data/data-linestring.wkt.csv
+++ b/test_data/data-linestring.wkt.csv
@@ -1,3 +1,4 @@
 "col","geometry"
 0,"LINESTRING (30 10, 10 30, 40 40)"
 1,"LINESTRING EMPTY"
+2,
diff --git a/test_data/data-multilinestring-encoding_native.parquet b/test_data/data-multilinestring-encoding_native.parquet
index 04e2e7c0f6de2b1dee9bcb3a2e47dca08c8559e1..d0b558dfd8e15e53852f94ba5b13df46ad2fee5f 100644
GIT binary patch
delta 424
zcmZ1?xI!>Az%j^Bltt7*)Wk=WLzF>?fkA<RftQhik(a}Ui2(wbp|mK2sEDY94p5zd
zD4Qsoj06WvMuG#Xgh7l!gM)cupp-oaBNOuz2DMj=GNLS!rVMNXk`m0x`8lFYVjN<x
z7{wkjszs=A;5K4%BeTlHa|VnO6aULFvP_m?)Zk)aWMWNGXJSxWFxijMm`6gag-PrW
zquLi{wI!2#8O4FB{xPdhJSQ^w0iz6(K5mdcm274PwK<%V{TcNHD}io+Y5u^jwt#(d
xAEOvV_vBMdm3ADGGLklwI9!zxj|(-J2sopTMT6rLs~CqGhuE#n>sY=q0|1c%O>6)F

delta 398
zcmZ1>xJ1x8z%j^Bltt7)RL4h@O_V{1fkA?SftS&SkpTjjptLB1sEDY94p5B%5VOfh
za4=1@7j=P2OK?EdfhaB&kT4q~6SE3~+9O68Q5H#41~vgn3FhSd98o4QHnCfbK%^F+
z#sM{lk!j+(5Js`dj7&0&Op_HDHMp4=nOIZQnHbb&uucwS(i5lv>Jt-dVG{epsP=|g
zZO-J0jN)LWqLcSAIw5Q20%=vrW@b>EGC6?Bm<Oio1-sge$^A?ST_Te&FqLbtNy<pt
h5bsJAW};jq#6-w;Ax^PJtYU0x9AcL?Z({kz3;@r+NeKV|

diff --git a/test_data/data-multilinestring-encoding_wkb.parquet b/test_data/data-multilinestring-encoding_wkb.parquet
index b3e7b26f1f88cf989d57c08facecd0c6fe5c1e74..0c84ec016587fa97da62b68548e8ae51342c27ee 100644
GIT binary patch
delta 365
zcmeC?>*os%a11gOWf64{HSrPU5M>ZzU{GLS;ALcB<mIqoVt@c<C@sn$Dk3VO15_s<
z$|lMtBf$Zak>G$TVGv``;9#B@C}q#V$i)1FLG2Zzj3|qwDFd5;qy%$vevT-U7>C#^
zMzKeXY7uH2c#K%Jf(vGk8Pnu`=2_+p3|uN8Co!`#F{?1BP2i9b<$;>ak)E2Ln_5y-
z2{!%;o7fu`wK=S6OPFQAde=<8$)s(^At@uNLxF=-G)b{@Eyq3%F%C5jv2&YUSdKFR
E0M5`i{{R30

delta 332
zcmeC@>*li#a11gOWf3(H)$tK!6J-!$V31&7;AOO7WPkuBC@sn$Dk3VO15_gb#B4GW
z988l9m_=P+@)8_Sg&>Mc1tiSI$i%F|p!SGSMwCU;lz~k^Qi3@-KSz{Fj7{toBM_+t
zsBypyVq}`QZUq<2bTfv@hD@`}(M;;#kP+p9n#7TwnxC6mQd9{x?h2dO8y2-GtZH+Z
nWx#r`Yp_YmNa~Paxr(M53Ffci6kEq3#-_$0c5Jf;%W)<EA}cg)

diff --git a/test_data/data-multilinestring.wkt.csv b/test_data/data-multilinestring.wkt.csv
index 17294fa..1c5f1ea 100644
--- a/test_data/data-multilinestring.wkt.csv
+++ b/test_data/data-multilinestring.wkt.csv
@@ -2,3 +2,4 @@
 0,"MULTILINESTRING ((30 10, 10 30, 40 40))"
 1,"MULTILINESTRING ((10 10, 20 20, 10 40), (40 40, 30 30, 40 20, 30 10))"
 2,"MULTILINESTRING EMPTY"
+3,
diff --git a/test_data/data-multipoint-encoding_native.parquet b/test_data/data-multipoint-encoding_native.parquet
index 78bd1329a1cebc68b456cccaa614804a2f7fb8ef..9bcbf13a1abd0755e7ca77d37d75d21f2ca941d5 100644
GIT binary patch
delta 427
zcmaFP|C&EEz%j^Bltt7*)Wk=WLzF>?fkA<RftQhik(a}Ui2(wbp|mK2sEDY94p5zd
zD4Qsoj06WvMuG#Xgh7l!gM)cupp-oaBNOuz2DMj=GNLS!rVMNXk`m0x`8lFYVjN<x
z7{wkjszs=A;5K5i4YSI`a|Voj6aULFvP_m|5uGf<B)~2PHgw8l4@Na!KCvH+VjmdQ
zzA&pTVV?YfQ4Fk3ed0Nh$!8g5&~$P@biU%4?8&GoPziJfO#3NzwO^CpGm1lWPhP?l
mWyc{YBWXj4(^VPqI+0n8kUKOuZn27SsBwrL+U(CVgBbuOW=`_}

delta 394
zcmaFO|D4}Cz%j^Bltt7)RL4h@O_V{1fkA?SftS&SkpTjjptLB1sEDY94p5B%5VOfh
za4=1@7j=P2OK?EdfhaB&kT4q~6SE3~+9O68Q5H#41~vgn3FhSd98o4QHnCfbK%^F+
z#sM{lk!i96v&!VXj3N`C$uROve#a;S6l!D=ovgqlz|O|R{DeWRXR;5Y8V`@y4@R*M
zjB0O~)#gmDV-yD~)117YQ4d`k2SnQ=j>*1^ih`9udtrJHv8#Pzn_SN*2G-oD!6qpq
jX+wfTRT=TPMV(oVpq1*JVy9Te*wi@0c5M!3nZXPI?FUTq

diff --git a/test_data/data-multipoint-encoding_wkb.parquet b/test_data/data-multipoint-encoding_wkb.parquet
index 6583b3e2531a1722560ea61740aa5e0ebd663c37..3d8e35d4b18d832da3260b9a9be426203274e2b4 100644
GIT binary patch
delta 329
zcmcb}bCoAFz%j^Bltt7*)Wk=WLzF>?fkA<RftQhik(a}Ui2(wbp|mK2sEDY94p5zd
zD4Qsoj06WvMuG#Xgh7l!gM)cupp-oaBNOuz2DMj=GNLS!rVMNXk`m0x`8lFYVjN<x
z7{wkjszs=A;4xxV4;RcHGp5N7jKv(x>`crm3~Ik7zhe>?;1IjSB6f{g?Foz85@s2&
ritNb^%tDjnnY8UVBxNLZD6m?E3BN7E9DCTrIMg`Ac5I%_yp0I}ZaOp6

delta 292
zcmcc0bCJh7z%j^Bltt7)RL4h@O_V{1fkA?SftS&SkpTjjptLB1sEDY94p5B%5VOfh
za4=1@7j=P2OK?EdfhaB&kT4q~6SE3~+9O68Q5H#41~vgn3FhSd98o4QHnCfbK%^F+
z#sM{lk!j+(9xj;KW(<=hnX@OaXXN038uDrKenxQtHnB@AV%M0}?y#uMVU_`_s-2w7
jq^-dwDI=*vyiF=hYB&w7<5<Hc#-_$0wqf&p=50&>MmaGd

diff --git a/test_data/data-multipoint.wkt.csv b/test_data/data-multipoint.wkt.csv
index d28b0a8..00d926f 100644
--- a/test_data/data-multipoint.wkt.csv
+++ b/test_data/data-multipoint.wkt.csv
@@ -2,3 +2,4 @@
 0,"MULTIPOINT ((30 10))"
 1,"MULTIPOINT ((10 40), (40 30), (20 20), (30 10))"
 2,"MULTIPOINT EMPTY"
+3,
diff --git a/test_data/data-multipolygon-encoding_native.parquet b/test_data/data-multipolygon-encoding_native.parquet
index 060beea13a3db5328d34911d52ce025a14000fa1..cc2b01f9d50d8121109f5f4c86ce2c2b6e8eb93d 100644
GIT binary patch
delta 423
zcmaDO^ijw;z%j^BltnZ^)W%1YOO!!~fkA_TftQhik(YykiI>BMnE?V=ptLB1sF<jf
z4p6OtD4Qsoj06YEM0-(Jn6v~3R2_)o)?fmXT#U@j9c~P2Ul?UXStLyv*aRdcn3MB!
zM4806#Qrdfy<t>KP~(6a#K<z4kws<VB|Ao&$%2eBK<))b(aFw?;=(M9%$yx;1sTuW
zddvbQFfgdyW1YN^$y%@iXpW88878p}Olq@O)YdUiW@Hru>nlJpqnJrtjD?YfQ$SHc
z&`80+Q8dgTsR-y-wPTYPFgftph<#%cJIAE<i$m?lBvx^-0h*I_nWt%RNy<ptkl=b%
cCW0<gXToEvI;Yq^b}=qB4zX98Jy=y(0BNdH(f|Me

delta 435
zcmew;^hU@&z%j^Bltt7*)Wk=WLzF>?fkA<RftQhik(a}Ui2(wbp|mK2sEDY94p5zd
zD4Qsoj06Ys<PXfEE--ls4yZy9#iat$&cVpU{DeX66{CzOi=-(7n}DPQb8>!;D3ch6
z*egb{M~rF_Y8)_w7?~%tv8YVEZpUc_vtjZ?7GWkarpXgnL??SNiVHC@GIMsY6=Xbf
z>oE(Mz`&q(WpW{-1HYBn878p}OllKY)Rr*INFa2nPyWN`iQTvgCUHif!Av4dd$=a+
zvRDgL0$pV#_KivG9Fy7y4z+iaKSETns82Rvo}s}ZDI;k^iceG-2{=xJu+5xeYuLp&
O)HuW*Z1!bUVF3Wkl~$Yp

diff --git a/test_data/data-multipolygon-encoding_wkb.parquet b/test_data/data-multipolygon-encoding_wkb.parquet
index 65a3cec73577ae336568c10043aaf909b93d52cd..77d5a5523c74fad00836f1ba7cbad60cb97a47f2 100644
GIT binary patch
delta 343
zcmcaF_(ae-z%j^BltnZ^)W%1YOO!!~fkA_TftQhik(YykiI>BMnE?V=ptLB1sF<jf
z4p6OtD4Qsoj06YEM0-(Jn6v~3R2_)o)?fmXT#U@j9c~P2Ul?UXStLyv*aRdcn3MB!
zM4806#Qrdfy<t>KP~(6a#K^L79Xlfz%ycuR&HEYU81<R?nV6q2s9oTe5#@mz!I7St
zpPO1zR0%fc7q8eN4z(w2YU`L~z?#GuC$R}lHeu1$;F6S))FIaPiR)A}Nwr3p;~T#i
Oml}sy-)3pHBTN83qdMOJ

delta 413
zcmaDNcwewSz%j^Bltt7*)Wk=WLzF>?fkA<RftQhik(a}Ui2(wbp|mK2sEDY94p5zd
zD4Qsoj06WvMuG#Xgh7l!1B)n7mP-X>90wy4^AiTOSBx^EERv=SY?cC&63of@IigHr
z9Ad8+#U3%LMW}HQWAbEq7MY1RA9BGQW5%#qkXek;9L=U<+%lp(Ae(>|aipi_=cbkv
zRf28%#VfXmL+uuu+7f0Nu;yS!4-QEgNgagiC~=&WrWy%$YcO((J>eJQP~#A5*{r~J
Ggb4t61vXUx

diff --git a/test_data/data-multipolygon.wkt.csv b/test_data/data-multipolygon.wkt.csv
index 79b971f..211a681 100644
--- a/test_data/data-multipolygon.wkt.csv
+++ b/test_data/data-multipolygon.wkt.csv
@@ -3,3 +3,4 @@
 1,"MULTIPOLYGON (((30 20, 45 40, 10 40, 30 20)), ((15 5, 40 10, 10 20, 5 10, 15 5)))"
 2,"MULTIPOLYGON (((40 40, 20 45, 45 30, 40 40)), ((20 35, 10 30, 10 10, 30 5, 45 20, 20 35), (30 20, 20 15, 20 25, 30 20)))"
 3,"MULTIPOLYGON EMPTY"
+4,
diff --git a/test_data/data-point-encoding_native.parquet b/test_data/data-point-encoding_native.parquet
index 774dbcced094ab0d1be1ba7526fbef64cd8199fd..f9d3fdf66aefa928c8dba74f6bc05f94da6f0668 100644
GIT binary patch
delta 695
zcmbQix0=sBz%j^Bltt7*)Wk=WLzF>?fkA<RftQhik(a}Ui2(wbp|mK2sEDY94p5zd
zD4Qsoj06Ys#C}m1n6v~3R2_)oQUU4aU}R!`!l3qwQAU(S(v*QsKvIG^IX_2~NsL46
z6{FZAMzshv4yZYd%#-gisni1<V<4*I1GE6-7zqXjRz?;(2Ud0sh9C71^IaS;9Ribq
zxxvmsj6nnH9v26MD2mIbFsS`wkrCyBxJ-j1JvBeqEVZPllCc8p!dr}Dw;0vlFsm(L
z2D%Zi6_dS~<!rGzU4>YuPc~!{^+h;U#R1)^OE}P+3U+BFl1nGCtDOSH3e=?_6B%75
xZ(x?2oXTXa!67LlX+uI(sJdeW4%uNNZ%DrD<vHiEigBoMh)rYNe1qA6832RKWWN9a

delta 594
zcmZ3@H-pbSz%j^BltolQRK*8KF$gg*2-q+{0V9+aWe^n-71IIA3y89bvdKtrFix}=
zb%sexaKJPGX-)+uAj!hW$jqd6gHcA5MbeakO+ZqDIXOQ^lu3+5>;$9O1x7UwH4dmT
zjEs{%GO1W%cY~b+#ONROm~Mc`*i9~E6qT`afH?xJ1KAOeCOa}pio@)^#H==n8E7+-
zW=6ZovCMMW*d3rk&;crNt5qDt7&M?JBZ~s<LUzzU_Q`e361p%Ky<<~b2MSV%i;(pp
tnKn6!DN=()Qby8-gg{kgB+l>hOk_Gqj!WzYix`U<hu8;}&99jqm;q`FP38aq

diff --git a/test_data/data-point-encoding_wkb.parquet b/test_data/data-point-encoding_wkb.parquet
index e9c5f7402cc7484d35afc98e463f3f15ee41ec59..e50d4b2ffc8dd58bc8c2b992fe8bbabb78d1f09d 100644
GIT binary patch
delta 365
zcmaFB^^MCuz%j^Bltt7*)Wk=WLzF>?fkA<RftQhik(a}Ui2(wbp|mK2sEDY94p5zd
zD4Qsoj06Ys#C}m1n6v~3R2_)oQUU4aU}R!`!l3qwQAU(S(v*QsKvIG^IX_2~NsL46
z6{FZAMzshv4yZYd%#$UVRD6MsnZ_s@;RCb*<Qi`YQJ^`Dj0}Q|EOri@Tr4UMR+0=X
zKk7Lpfdr7L;^F{Q00b@$=<YLPnw-z5F9354GdB~n3WM5?$@Wa*0vuxJn8c<qsm)?h
oTf!^@R+*#0At@uNLxO8n8F3os$SJmiMT|p@L+sn;4NM!D01SIKR{#J2

delta 352
zcmeyy^?=Jfz%j^BltolQRK*8KF$gg*2-q+{0V9+aWe^n-71IIA3y89bvdKtrFix}=
zb%sexaKJPGX-)+uAj!hW$jqd6gHcA5MbeakO+ZqDIXOQ^lu3+5>;$9O1x7UwH4dmT
zjEs}@nN%EshNp-+!Q5aZAqq5wk&!`=k;Tq|lZ!>g!Ag=r<wrflrXTg_PBLSdEYE1j
zi)?$(<Tyrg0T!_tOk%$n)jlw*O=6Y-E7PC6nn`H#F-C0-7D*XN9TJ?W%81ucVU7pP
QVk~MLVy8AAW7@z30QIpo(f|Me

diff --git a/test_data/data-point.wkt.csv b/test_data/data-point.wkt.csv
index 6f8b211..8164a4b 100644
--- a/test_data/data-point.wkt.csv
+++ b/test_data/data-point.wkt.csv
@@ -1,3 +1,5 @@
 "col","geometry"
 0,"POINT (30 10)"
 1,"POINT EMPTY"
+2,
+3,"POINT (40 40)"
diff --git a/test_data/data-polygon-encoding_native.parquet b/test_data/data-polygon-encoding_native.parquet
index 5671f55777273ccf465eb39ee387fb75aa743bfd..344168ee764ff558e48fc8362f419b7353f6a2f7 100644
GIT binary patch
delta 423
zcmeAdoGcg`;22~m$|C9@YT_fxA<7`cz@Wguz{|+M$jf2F!~g-zP+F8hR76xl2dGX!
zlueXPMuGz-Bf$Yx!XU<=!NEK+P|BWzk%{>UgW4-b8BrEVQwBBxNeSlU{2WmxF%Gd;
zjAD-%)gsh5a2qjs4ztR{bGD2MlX)0r7+EH3GHUR!Ffy}tuw7fs!Kk)xax9}UkAm0^
zCb3yeYJDtfOC~e2hy&H_W73>_m(c=QA1_Fs2a67qHj~;V&dG6%dIFU|XDEoRU=o|c
wq_&6yXj&sw^8z~#Nf}8SN}R3Ah{tj2%mkgF&M9_^O^id0Lu|t4lPt%W0nIo^uK)l5

delta 384
zcmbO%*e_@u;22~m$|7nYs^cTdCdwehz#zfEz{_aE$N&LMP+F8hR76xl2dG8>h}mQ$
zIG85di@Lz1B{-n!Kopk>NSKY0iCKj~?GdAlD2t>i1Dk-P1aoqJjwq8Do7gQzAW{oZ
z<A9pO$TWEYv&zIvwv2L<1sG+3TwO*D9wtU+)(*C7i#Zt8woOiCH0F^L+rcC@i%G4G
zMQzUH1&rcAb+Rn#lkYQHVd(Q<(P7eNQad#{o5`40PHY8}*c2wUSsZHH*e5S!6oYDf
q$yBMqCMhFnLxL++8S%J9gAuodb6GSvcCm@Esd0#PY(C3!j2QqPY(^CT

diff --git a/test_data/data-polygon-encoding_wkb.parquet b/test_data/data-polygon-encoding_wkb.parquet
index 06f065d1dcfc85d494a3bd3d37412511f5f2b0e0..45d2cba07e53d7bbf0616f9a879900142cd5e0b8 100644
GIT binary patch
delta 326
zcmX@acZ@GIz%j^Bltt7*)Wk=WLzF>?fkA<RftQhik(a}Ui2(wbp|mK2sEDY94p5zd
zD4Qsoj06WvMuG#Xgh7l!gM)cupp-oaBNOuz2DMj=GNLS!rVMNXk`m0x`8lFYVjN<x
z7{wkjszs=A;4xy=CN7vgW=xZPnKlS8FmS1WoW#t|#H_-gc46{7W^n-yu}AD;YgpAb
qu&OO#mI14LYsVoeBdJ4y)hbN5O+Umbc7{`oLybeMZ}S9}O-ullc`#T2

delta 329
zcmX@ccZknAz%j^Bltt7)RL4h@O_V{1fkA?SftS&SkpTjjptLB1sEDY94p5B%5VOfh
za4=1@7j=P2OK?EdfhaB&kT4q~6SE3~+9O68Q5H#41~vgn3FhSd98o4QHnCfbK%^F+
z#sM{lk!j+(O<XXu%@`&JGHozNGw1||j3^J(Add9Z{M^)%qDru7kJ!c5u&S+KRhz>s
n1J?UigH2LKQipiERW#K|F#Zsy*d9(XHZ=~hw#`#mHZcJJ_E9zS

diff --git a/test_data/data-polygon.wkt.csv b/test_data/data-polygon.wkt.csv
index c528373..0574892 100644
--- a/test_data/data-polygon.wkt.csv
+++ b/test_data/data-polygon.wkt.csv
@@ -2,3 +2,4 @@
 0,"POLYGON ((30 10, 40 40, 20 40, 10 20, 30 10))"
 1,"POLYGON ((35 10, 45 45, 15 40, 10 20, 35 10), (20 30, 35 35, 30 20, 20 30))"
 2,"POLYGON EMPTY"
+3,
diff --git a/test_data/generate_test_data.py b/test_data/generate_test_data.py
index 8fe7377..eafb6d9 100644
--- a/test_data/generate_test_data.py
+++ b/test_data/generate_test_data.py
@@ -64,6 +64,8 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
 geometries_wkt = [
     "POINT (30 10)",
     "POINT EMPTY",
+    None,
+    "POINT (40 40)",
 ]
 
 point_type = pa.struct(
@@ -72,7 +74,7 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
         pa.field("y", pa.float64(), nullable=False)
     ]
 )
-geometries = pa.array([(30, 10), (float("nan"), float("nan"))], type=point_type)
+geometries = pa.array([(30, 10), (float("nan"), float("nan")), (float("nan"), float("nan")), (40, 40)], type=point_type)
 
 write_encoding_files(
     geometries_wkt, geometries, geometry_type="Point"
@@ -83,11 +85,12 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
 geometries_wkt = [
     "LINESTRING (30 10, 10 30, 40 40)",
     "LINESTRING EMPTY",
+    None
 ]
 
 linestring_type = pa.list_(pa.field("vertices", point_type, nullable=False))
 geometries = pa.array(
-    [[(30, 10), (10, 30), (40, 40)], []], type=linestring_type)
+    [[(30, 10), (10, 30), (40, 40)], [], []], type=linestring_type)
 
 write_encoding_files(
     geometries_wkt, geometries, geometry_type="LineString"
@@ -99,6 +102,7 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
     "POLYGON ((30 10, 40 40, 20 40, 10 20, 30 10))",
     "POLYGON ((35 10, 45 45, 15 40, 10 20, 35 10), (20 30, 35 35, 30 20, 20 30))",
     "POLYGON EMPTY",
+    None,
 ]
 
 polygon_type = pa.list_(
@@ -108,9 +112,10 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
 )
 geometries = pa.array(
     [
-        [[(30.0, 10.0), (40.0, 40.0), (20.0, 40.0), (10.0, 20.0), (30.0, 10.0)]],
-        [[(35.0, 10.0), (45.0, 45.0), (15.0, 40.0), (10.0, 20.0), (35.0, 10.0)],
-         [(20.0, 30.0), (35.0, 35.0), (30.0, 20.0), (20.0, 30.0)]],
+        [[(30, 10), (40, 40), (20, 40), (10, 20), (30, 10)]],
+        [[(35, 10), (45, 45), (15, 40), (10, 20), (35, 10)],
+         [(20, 30), (35, 35), (30, 20), (20, 30)]],
+        [],
         [],
     ],
     type=polygon_type
@@ -126,14 +131,16 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
     "MULTIPOINT ((30 10))",
     "MULTIPOINT ((10 40), (40 30), (20 20), (30 10))",
     "MULTIPOINT EMPTY",
+    None,
 ]
 
 multipoint_type = pa.list_(pa.field("points", point_type, nullable=False))
 geometries = pa.array(
     [
-        [(30.0, 10.0)],
-        [(10.0, 40.0), (40.0, 30.0), (20.0, 20.0), (30.0, 10.0)],
-        []
+        [(30, 10)],
+        [(10, 40), (40, 30), (20, 20), (30, 10)],
+        [],
+        [],
     ],
     type=multipoint_type
 )
@@ -148,6 +155,7 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
     "MULTILINESTRING ((30 10, 10 30, 40 40))",
     "MULTILINESTRING ((10 10, 20 20, 10 40), (40 40, 30 30, 40 20, 30 10))",
     "MULTILINESTRING EMPTY",
+    None,
 ]
 
 multilinestring_type = pa.list_(
@@ -155,9 +163,10 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
 )
 geometries = pa.array(
     [
-        [[(30.0, 10.0), (10.0, 30.0), (40.0, 40.0)]],
-        [[(10.0, 10.0), (20.0, 20.0), (10.0, 40.0)],
-         [(40.0, 40.0), (30.0, 30.0), (40.0, 20.0), (30.0, 10.0)]],
+        [[(30, 10), (10, 30), (40, 40)]],
+        [[(10, 10), (20, 20), (10, 40)],
+         [(40, 40), (30, 30), (40, 20), (30, 10)]],
+        [],
         [],
     ],
     type=multilinestring_type
@@ -174,17 +183,19 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
     "MULTIPOLYGON (((30 20, 45 40, 10 40, 30 20)), ((15 5, 40 10, 10 20, 5 10, 15 5)))",
     "MULTIPOLYGON (((40 40, 20 45, 45 30, 40 40)), ((20 35, 10 30, 10 10, 30 5, 45 20, 20 35), (30 20, 20 15, 20 25, 30 20)))",
     "MULTIPOLYGON EMPTY",
+    None,
 ]
 
 multipolygon_type = pa.list_(pa.field("polygons", polygon_type, nullable=False))
 geometries = pa.array(
     [
-        [[[(30.0, 10.0), (40.0, 40.0), (20.0, 40.0), (10.0, 20.0), (30.0, 10.0)]]],
-        [[[(30.0, 20.0), (45.0, 40.0), (10.0, 40.0), (30.0, 20.0)]],
-         [[(15.0, 5.0), (40.0, 10.0), (10.0, 20.0), (5.0, 10.0), (15.0, 5.0)]]],
-        [[[(40.0, 40.0), (20.0, 45.0), (45.0, 30.0), (40.0, 40.0)]],
-         [[(20.0, 35.0), (10.0, 30.0), (10.0, 10.0), (30.0, 5.0), (45.0, 20.0), (20.0, 35.0)],
-          [(30.0, 20.0), (20.0, 15.0), (20.0, 25.0), (30.0, 20.0)]]],
+        [[[(30, 10), (40, 40), (20, 40), (10, 20), (30, 10)]]],
+        [[[(30, 20), (45, 40), (10, 40), (30, 20)]],
+         [[(15, 5), (40, 10), (10, 20), (5, 10), (15, 5)]]],
+        [[[(40, 40), (20, 45), (45, 30), (40, 40)]],
+         [[(20, 35), (10, 30), (10, 10), (30, 5), (45, 20), (20, 35)],
+          [(30, 20), (20, 15), (20, 25), (30, 20)]]],
+        [],
         [],
     ],
     type=multipolygon_type

From 2981b6000da33a2f3f4f639479c88003fe6ba172 Mon Sep 17 00:00:00 2001
From: Joris Van den Bossche <jorisvandenbossche@gmail.com>
Date: Thu, 2 May 2024 16:33:50 +0200
Subject: [PATCH 6/8] rename csv files

---
 test_data/data-linestring.wkt.csv      | 4 ----
 test_data/data-multilinestring.wkt.csv | 5 -----
 test_data/data-multipoint.wkt.csv      | 5 -----
 test_data/data-multipolygon.wkt.csv    | 6 ------
 test_data/data-point.wkt.csv           | 5 -----
 test_data/data-polygon.wkt.csv         | 5 -----
 test_data/generate_test_data.py        | 2 +-
 7 files changed, 1 insertion(+), 31 deletions(-)
 delete mode 100644 test_data/data-linestring.wkt.csv
 delete mode 100644 test_data/data-multilinestring.wkt.csv
 delete mode 100644 test_data/data-multipoint.wkt.csv
 delete mode 100644 test_data/data-multipolygon.wkt.csv
 delete mode 100644 test_data/data-point.wkt.csv
 delete mode 100644 test_data/data-polygon.wkt.csv

diff --git a/test_data/data-linestring.wkt.csv b/test_data/data-linestring.wkt.csv
deleted file mode 100644
index a3a320d..0000000
--- a/test_data/data-linestring.wkt.csv
+++ /dev/null
@@ -1,4 +0,0 @@
-"col","geometry"
-0,"LINESTRING (30 10, 10 30, 40 40)"
-1,"LINESTRING EMPTY"
-2,
diff --git a/test_data/data-multilinestring.wkt.csv b/test_data/data-multilinestring.wkt.csv
deleted file mode 100644
index 1c5f1ea..0000000
--- a/test_data/data-multilinestring.wkt.csv
+++ /dev/null
@@ -1,5 +0,0 @@
-"col","geometry"
-0,"MULTILINESTRING ((30 10, 10 30, 40 40))"
-1,"MULTILINESTRING ((10 10, 20 20, 10 40), (40 40, 30 30, 40 20, 30 10))"
-2,"MULTILINESTRING EMPTY"
-3,
diff --git a/test_data/data-multipoint.wkt.csv b/test_data/data-multipoint.wkt.csv
deleted file mode 100644
index 00d926f..0000000
--- a/test_data/data-multipoint.wkt.csv
+++ /dev/null
@@ -1,5 +0,0 @@
-"col","geometry"
-0,"MULTIPOINT ((30 10))"
-1,"MULTIPOINT ((10 40), (40 30), (20 20), (30 10))"
-2,"MULTIPOINT EMPTY"
-3,
diff --git a/test_data/data-multipolygon.wkt.csv b/test_data/data-multipolygon.wkt.csv
deleted file mode 100644
index 211a681..0000000
--- a/test_data/data-multipolygon.wkt.csv
+++ /dev/null
@@ -1,6 +0,0 @@
-"col","geometry"
-0,"MULTIPOLYGON (((30 10, 40 40, 20 40, 10 20, 30 10)))"
-1,"MULTIPOLYGON (((30 20, 45 40, 10 40, 30 20)), ((15 5, 40 10, 10 20, 5 10, 15 5)))"
-2,"MULTIPOLYGON (((40 40, 20 45, 45 30, 40 40)), ((20 35, 10 30, 10 10, 30 5, 45 20, 20 35), (30 20, 20 15, 20 25, 30 20)))"
-3,"MULTIPOLYGON EMPTY"
-4,
diff --git a/test_data/data-point.wkt.csv b/test_data/data-point.wkt.csv
deleted file mode 100644
index 8164a4b..0000000
--- a/test_data/data-point.wkt.csv
+++ /dev/null
@@ -1,5 +0,0 @@
-"col","geometry"
-0,"POINT (30 10)"
-1,"POINT EMPTY"
-2,
-3,"POINT (40 40)"
diff --git a/test_data/data-polygon.wkt.csv b/test_data/data-polygon.wkt.csv
deleted file mode 100644
index 0574892..0000000
--- a/test_data/data-polygon.wkt.csv
+++ /dev/null
@@ -1,5 +0,0 @@
-"col","geometry"
-0,"POLYGON ((30 10, 40 40, 20 40, 10 20, 30 10))"
-1,"POLYGON ((35 10, 45 45, 15 40, 10 20, 35 10), (20 30, 35 35, 30 20, 20 30))"
-2,"POLYGON EMPTY"
-3,
diff --git a/test_data/generate_test_data.py b/test_data/generate_test_data.py
index eafb6d9..f04731c 100644
--- a/test_data/generate_test_data.py
+++ b/test_data/generate_test_data.py
@@ -39,7 +39,7 @@
 def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
 
     table = pa.table({"col": range(len(geometries_wkt)), "geometry": geometries_wkt})
-    write_csv(table, HERE / f"data-{geometry_type.lower()}.wkt.csv")
+    write_csv(table, HERE / f"data-{geometry_type.lower()}-wkt.csv")
 
     # WKB encoding
     table = pa.table(

From 6c9390fd05e7d7db642ab520328d5e64c14d070b Mon Sep 17 00:00:00 2001
From: Joris Van den Bossche <jorisvandenbossche@gmail.com>
Date: Thu, 2 May 2024 16:37:33 +0200
Subject: [PATCH 7/8] add back csv files

---
 test_data/data-linestring-wkt.csv      | 4 ++++
 test_data/data-multilinestring-wkt.csv | 5 +++++
 test_data/data-multipoint-wkt.csv      | 5 +++++
 test_data/data-multipolygon-wkt.csv    | 6 ++++++
 test_data/data-point-wkt.csv           | 5 +++++
 test_data/data-polygon-wkt.csv         | 5 +++++
 6 files changed, 30 insertions(+)
 create mode 100644 test_data/data-linestring-wkt.csv
 create mode 100644 test_data/data-multilinestring-wkt.csv
 create mode 100644 test_data/data-multipoint-wkt.csv
 create mode 100644 test_data/data-multipolygon-wkt.csv
 create mode 100644 test_data/data-point-wkt.csv
 create mode 100644 test_data/data-polygon-wkt.csv

diff --git a/test_data/data-linestring-wkt.csv b/test_data/data-linestring-wkt.csv
new file mode 100644
index 0000000..a3a320d
--- /dev/null
+++ b/test_data/data-linestring-wkt.csv
@@ -0,0 +1,4 @@
+"col","geometry"
+0,"LINESTRING (30 10, 10 30, 40 40)"
+1,"LINESTRING EMPTY"
+2,
diff --git a/test_data/data-multilinestring-wkt.csv b/test_data/data-multilinestring-wkt.csv
new file mode 100644
index 0000000..1c5f1ea
--- /dev/null
+++ b/test_data/data-multilinestring-wkt.csv
@@ -0,0 +1,5 @@
+"col","geometry"
+0,"MULTILINESTRING ((30 10, 10 30, 40 40))"
+1,"MULTILINESTRING ((10 10, 20 20, 10 40), (40 40, 30 30, 40 20, 30 10))"
+2,"MULTILINESTRING EMPTY"
+3,
diff --git a/test_data/data-multipoint-wkt.csv b/test_data/data-multipoint-wkt.csv
new file mode 100644
index 0000000..00d926f
--- /dev/null
+++ b/test_data/data-multipoint-wkt.csv
@@ -0,0 +1,5 @@
+"col","geometry"
+0,"MULTIPOINT ((30 10))"
+1,"MULTIPOINT ((10 40), (40 30), (20 20), (30 10))"
+2,"MULTIPOINT EMPTY"
+3,
diff --git a/test_data/data-multipolygon-wkt.csv b/test_data/data-multipolygon-wkt.csv
new file mode 100644
index 0000000..211a681
--- /dev/null
+++ b/test_data/data-multipolygon-wkt.csv
@@ -0,0 +1,6 @@
+"col","geometry"
+0,"MULTIPOLYGON (((30 10, 40 40, 20 40, 10 20, 30 10)))"
+1,"MULTIPOLYGON (((30 20, 45 40, 10 40, 30 20)), ((15 5, 40 10, 10 20, 5 10, 15 5)))"
+2,"MULTIPOLYGON (((40 40, 20 45, 45 30, 40 40)), ((20 35, 10 30, 10 10, 30 5, 45 20, 20 35), (30 20, 20 15, 20 25, 30 20)))"
+3,"MULTIPOLYGON EMPTY"
+4,
diff --git a/test_data/data-point-wkt.csv b/test_data/data-point-wkt.csv
new file mode 100644
index 0000000..8164a4b
--- /dev/null
+++ b/test_data/data-point-wkt.csv
@@ -0,0 +1,5 @@
+"col","geometry"
+0,"POINT (30 10)"
+1,"POINT EMPTY"
+2,
+3,"POINT (40 40)"
diff --git a/test_data/data-polygon-wkt.csv b/test_data/data-polygon-wkt.csv
new file mode 100644
index 0000000..0574892
--- /dev/null
+++ b/test_data/data-polygon-wkt.csv
@@ -0,0 +1,5 @@
+"col","geometry"
+0,"POLYGON ((30 10, 40 40, 20 40, 10 20, 30 10))"
+1,"POLYGON ((35 10, 45 45, 15 40, 10 20, 35 10), (20 30, 35 35, 30 20, 20 30))"
+2,"POLYGON EMPTY"
+3,

From 3466e1ff37cf7d1577dee943eb0199f3fb08b128 Mon Sep 17 00:00:00 2001
From: Joris Van den Bossche <jorisvandenbossche@gmail.com>
Date: Wed, 8 May 2024 09:16:18 +0200
Subject: [PATCH 8/8] properly specify mask when creating the Arrow data

---
 .../data-linestring-encoding_native.parquet   | Bin 2023 -> 2023 bytes
 .../data-linestring-encoding_wkb.parquet      | Bin 1474 -> 1474 bytes
 ...ta-multilinestring-encoding_native.parquet | Bin 2216 -> 2218 bytes
 .../data-multilinestring-encoding_wkb.parquet | Bin 1807 -> 1805 bytes
 .../data-multipoint-encoding_native.parquet   | Bin 2027 -> 2027 bytes
 .../data-multipoint-encoding_wkb.parquet      | Bin 1621 -> 1622 bytes
 .../data-multipolygon-encoding_native.parquet | Bin 2417 -> 2421 bytes
 .../data-multipolygon-encoding_wkb.parquet    | Bin 2276 -> 2276 bytes
 test_data/data-point-encoding_native.parquet  | Bin 1835 -> 1835 bytes
 test_data/data-point-encoding_wkb.parquet     | Bin 1398 -> 1398 bytes
 .../data-polygon-encoding_native.parquet      | Bin 2195 -> 2197 bytes
 test_data/data-polygon-encoding_wkb.parquet   | Bin 1862 -> 1861 bytes
 test_data/generate_test_data.py               |  16 ++++++++++++++--
 13 files changed, 14 insertions(+), 2 deletions(-)

diff --git a/test_data/data-linestring-encoding_native.parquet b/test_data/data-linestring-encoding_native.parquet
index 32653ef94bb29fab1ec39a36a082423323fa53cf..264705df6fba8cd34a5d101feb41f118b5f76a22 100644
GIT binary patch
delta 49
zcmaFP|D1n=Eh8htWIM)ZEDTJ{Dw9t!mu#NHWX39LBxxhVzyLuE&a4a!0gge207B&o
AZ2$lO

delta 49
zcmaFP|D1n=Eh8i2WIM)ZER0OdDw9t!mu#NHWX39LAZa7RzyLuE&a4a!0gge207C}~
AZ2$lO

diff --git a/test_data/data-linestring-encoding_wkb.parquet b/test_data/data-linestring-encoding_wkb.parquet
index 276e4184eda51af7ef65e67abb4c7d54ba3e88a1..fd8b1d47c38c71228495b1426dd2cd850c270c43 100644
GIT binary patch
delta 26
gcmX@aeTaL*238>>NgWvm1|VWM#>~JF;22~G09#fCo&W#<

delta 26
gcmX@aeTaL*238>hNgWvm1|VWM#>~JF;22~G09!HzoB#j-

diff --git a/test_data/data-multilinestring-encoding_native.parquet b/test_data/data-multilinestring-encoding_native.parquet
index d0b558dfd8e15e53852f94ba5b13df46ad2fee5f..2bb58225a1020186e7b0eea4e0cd3867bb0223d6 100644
GIT binary patch
delta 199
zcmZ1>xJq!sUuFYQlSz!`vJx?@3=9mc8nd_<KnxLPW<~}k))aLn2DL@3lLMLbxGF@M
z#3aPpn8f}~?q)22>XX5)PbHg~L2be0047r|kOoGzPwZ-oHup10GcsW_&6Np5@#KCc
mdrq-$tYRE$9Afu2Z(@1KENUcaBg4P|K@69G&J1u2G6Vpg@G5Qq

delta 183
zcmZ1_xI%EkUuHc~qe+bB(&7=U3=9mc8nd_<Sbz)>W)?;!))aLn2DJr~*D{;3v@nVN
zncTxzfLFarHZz0ToXNT@rd%NPjA|d))fQ}?$Ry3kh+>K>6NakE6PfHe#Xhl$aj0>K
c-P*i`<sq}EfuxNL0|NvxTmrf;z%j@W02l%&`~Uy|

diff --git a/test_data/data-multilinestring-encoding_wkb.parquet b/test_data/data-multilinestring-encoding_wkb.parquet
index 0c84ec016587fa97da62b68548e8ae51342c27ee..08f9f0d2d0fe6c6d8525f32c00119808fe517e6f 100644
GIT binary patch
delta 91
zcmeC@>*bqpgYnA5oBC`#E)Lxcyp#DEwO9l_1tliiF`j1ZnViEc!}5YfZN}st=FOAu
jGWDSf?Pff+*@I;(laP_5jtm0>5HV=6Ffar-1{neXD>xb!

delta 93
zcmeC>>*t$rgYm}1oBHe=E)Ly{wv3bc7`0d#Oso_bCfhKcW}GlNi&=){4U5{G$z9Bw
lC*Nl3LlxT1cy6;B%T^{K14$hj1_mHv&|+a=2yhHC1OQC$8ejkb

diff --git a/test_data/data-multipoint-encoding_native.parquet b/test_data/data-multipoint-encoding_native.parquet
index 9bcbf13a1abd0755e7ca77d37d75d21f2ca941d5..b0435ef6c99a7a91617e510dd3a22156acd2a695 100644
GIT binary patch
delta 42
tcmaFO|C)b;D<dP@WH-iFjBJ|?nU=DO8cEv7Ffc$6gEcDyLx5wDApq-E2$28)

delta 42
tcmaFO|C)b;D<h-WWH-iFjAEM&nU=DO8c5p6Ffc$6gEcDyLx5wDApq|T2(JJD

diff --git a/test_data/data-multipoint-encoding_wkb.parquet b/test_data/data-multipoint-encoding_wkb.parquet
index 3d8e35d4b18d832da3260b9a9be426203274e2b4..8454f250b3a4211286d8ca5372d06094e13aec1d 100644
GIT binary patch
delta 121
zcmcc0bB$-h4aP+iZ>nlADlxLiDlw?=DzUgY$ci$!@QO0n>4`Ehm;f0jAch^UsD+Bd
zWL3tQjDIHYXOv;N!L0Ud^L<7kMkM~^`;0A|Vtd%cIMg`Ac5R-|yo5=}NK!|JfdPmZ
NUNJK;1ULp60stkb9svLV

delta 93
zcmcb{bCqYp4aNl%Z>qBLurM$%Ok$L0mP|C5tj0K#@z>-7j4~|OnAM(ae!wWih{T`#
ofU$*BY!90lhZ=|2j?D|0moNz#Nb1NiFaQz5D`p0U0LLIh0DILN9RL6T

diff --git a/test_data/data-multipolygon-encoding_native.parquet b/test_data/data-multipolygon-encoding_native.parquet
index cc2b01f9d50d8121109f5f4c86ce2c2b6e8eb93d..43a8d2f9dbc693371bc448bd14e87a509c7d3c27 100644
GIT binary patch
delta 281
zcmew;^i^ns03)B5sGp9gji`Vqn<$%%#N>%A!jt=0#I+e1R3~sUFfefb`TYI-{r~?M
z*nk`h7G_2U24>C<wt|djZaroJ6BroO9!)M}bmG0hB({Y~Z5E5#I%c>jnv?%9y6^>w
zhGDTlnt?%c0S*f$Ph=8iI>$9xm&ID3Qj|%|M(h`p*d->le;jImHh*NYVr0Z^rxq@a
plRq-qbBY~f7voan5c{;*mzAGI)JW1shJgWs7%YK43~&rG1OOKUN+<vT

delta 239
zcmew=^igPo03(l^sF#kYji`Vqn<(4l1B^PF3=GO0oD2*MoPR!lKY#!KKL!>c$AX20
zk(slDtsvu>TaQ`51O^7Rdy@@VoLJ5<iEWs?ld+S>Pc#U{96bgG^%+>rVc`@|R1h>$
zFmMzNGe|1Z0Gf4-YqBAWwP2+vlbDUzHzu)jOlrS4)PAr79po_CjQJ-cie;|M_%zsY
ZitS?;<5J@gd$l=`m7j&tU~(d-2ms!TLG1tl

diff --git a/test_data/data-multipolygon-encoding_wkb.parquet b/test_data/data-multipolygon-encoding_wkb.parquet
index 77d5a5523c74fad00836f1ba7cbad60cb97a47f2..538b8f4e64705c94988a8fc10636f8d77d17ad68 100644
GIT binary patch
delta 26
gcmaDN_(X8SB@Q7YNgWvm1|VWs%)-DB;22~G0A&va@c;k-

delta 26
gcmaDN_(X8SB@Q72NgWvm1|VWs%)-DB;22~G0A%Y0?*IS*

diff --git a/test_data/data-point-encoding_native.parquet b/test_data/data-point-encoding_native.parquet
index f9d3fdf66aefa928c8dba74f6bc05f94da6f0668..4e6489aa576363b2ebacf69ac49763a5fcc0c8e4 100644
GIT binary patch
delta 103
zcmZ3@x0-K*G$SL^WEn<TR%UJ{W|hfyjP^i&9b*BI3}h1JWnkb^VParlfXE6m0cAG_
qGBJTP)iJ71u4c9aD^%b7klB=3)JW1shJgWs7*?||Fa$UT83F*Ewh>_f

delta 103
zcmZ3@x0-K*G$SL!WEn<TRt`oc<|&iy80~@lI>rJZ8OS8c%fP^;!o<M90Ff1B0?KX<
qWMTqos$*22T+M6;R;a%DA+ss7sDY%73<CoMF|1}`U<hywG6VqC7!oM}

diff --git a/test_data/data-point-encoding_wkb.parquet b/test_data/data-point-encoding_wkb.parquet
index e50d4b2ffc8dd58bc8c2b992fe8bbabb78d1f09d..94b87d16e1b3d66953072f9f0699774a759e76e5 100644
GIT binary patch
delta 26
gcmeyy^^I#o3agNjq>c;&0}wH+U}j(la11g80A3;mng9R*

delta 26
gcmeyy^^I#o3agNTq>c;&0}wH+U}j(la11g80A2nCm;e9(

diff --git a/test_data/data-polygon-encoding_native.parquet b/test_data/data-polygon-encoding_native.parquet
index 344168ee764ff558e48fc8362f419b7353f6a2f7..68d1aa67ca4f3d1b29c4205a4b0eedf68eb2f02e 100644
GIT binary patch
delta 200
zcmbO%I8|^17bCNYsKsPnMsqo-6jlZX2G$jeW^sWSGR(}349u(@Y}XcZFsdDxoXBX(
zvWrP<&gB2hu9NRGniA0G!J@;Y&7^j1ayFAG_bMi_X-sNMIMfcXPhQCAF!?ppEk+a*
q-59YcwBr;z$0o+1#vwLk^I4XS%%Vn;HZnk$ff2(Zpeq9$gA4&#P%q2?

delta 204
zcmbO#I9YH57bCNgsM%y*MsqpI1XczH2G$jeW^u6q88XZ)jLfVZY}XcZFskiiot(#{
z$5kQ9B&Hy?gGp@GWJVU($qyJ!VY+3p>h@sKVbW$&yEHkM$&`Bqlh_m{wM86i``9Nd
uGU-e{$0#!SEz>O~Y-YPL;?g6;DRzoYj6;n>Y{KUAEE|~_4JJ2ohyVc2xi9ko

diff --git a/test_data/data-polygon-encoding_wkb.parquet b/test_data/data-polygon-encoding_wkb.parquet
index 45d2cba07e53d7bbf0616f9a879900142cd5e0b8..cce77baad2acf019d9d89d6509076f14996e6692 100644
GIT binary patch
delta 92
zcmX@cca(3!4aQd!Z<=zkEM(wKWMN{EP24QWD8R^gZn74W49hB3wRM}ln8K0ylf9Tu
ma*Cbd6ys3i5bN1Ig=GPgkddU03<CoYF=(?eFa$UT83F*{?-xw~

delta 88
zcmX@gcZ_es4aPSUZ<=y5EM$yfOk`lvOJvwA!6?AUcwzEFW*L?>tZEx3A7nl`S%|q0
gRp=sP-{#3I3z&oqBz0sM7=VaDn}vZPz%j@W00|Zu&Hw-a

diff --git a/test_data/generate_test_data.py b/test_data/generate_test_data.py
index f04731c..0e50a14 100644
--- a/test_data/generate_test_data.py
+++ b/test_data/generate_test_data.py
@@ -12,6 +12,7 @@
 import pathlib
 import copy
 
+import numpy as np
 import pyarrow as pa
 import pyarrow.parquet as pq
 from pyarrow.csv import write_csv
@@ -74,7 +75,11 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
         pa.field("y", pa.float64(), nullable=False)
     ]
 )
-geometries = pa.array([(30, 10), (float("nan"), float("nan")), (float("nan"), float("nan")), (40, 40)], type=point_type)
+geometries = pa.array(
+    [(30, 10), (float("nan"), float("nan")), (float("nan"), float("nan")), (40, 40)],
+    mask=np.array([False, False, True, False]),
+    type=point_type
+)
 
 write_encoding_files(
     geometries_wkt, geometries, geometry_type="Point"
@@ -90,7 +95,10 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
 
 linestring_type = pa.list_(pa.field("vertices", point_type, nullable=False))
 geometries = pa.array(
-    [[(30, 10), (10, 30), (40, 40)], [], []], type=linestring_type)
+    [[(30, 10), (10, 30), (40, 40)], [], []],
+    mask=np.array([False, False, True]),
+    type=linestring_type
+)
 
 write_encoding_files(
     geometries_wkt, geometries, geometry_type="LineString"
@@ -118,6 +126,7 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
         [],
         [],
     ],
+    mask=np.array([False, False, False, True]),
     type=polygon_type
 )
 
@@ -142,6 +151,7 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
         [],
         [],
     ],
+    mask=np.array([False, False, False, True]),
     type=multipoint_type
 )
 
@@ -169,6 +179,7 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
         [],
         [],
     ],
+    mask=np.array([False, False, False, True]),
     type=multilinestring_type
 )
 
@@ -198,6 +209,7 @@ def write_encoding_files(geometries_wkt, geometries_geoarrow, geometry_type):
         [],
         [],
     ],
+    mask=np.array([False, False, False, False, True]),
     type=multipolygon_type
 )