From 003d4fb1322b0988213e41b2258acc9482f93469 Mon Sep 17 00:00:00 2001 From: Ben Schneider Date: Sat, 9 Mar 2024 11:30:25 -0500 Subject: [PATCH] Mention the new `degf` parameter in the database vignette, since this parameter is important for large datasets. --- vignettes/srvyr-database.Rmd | 2 ++ 1 file changed, 2 insertions(+) diff --git a/vignettes/srvyr-database.Rmd b/vignettes/srvyr-database.Rmd index 39b51f2..1a3aa3c 100644 --- a/vignettes/srvyr-database.Rmd +++ b/vignettes/srvyr-database.Rmd @@ -187,6 +187,8 @@ it will save a lot space. object.size(acs_m_db_svy) ``` +For very large survey data sets with replicate weights, it is strongly recommended to specify a value in the `degf` parameter of `as_survey_rep()`: otherwise, the survey package will attempt to automatically determine the design degrees of freedom using a process that can be very slow for large data sets. The value to use for `degf` will often be specified in survey data documentation, but a common rule of thumb is the number of columns of replicate weights. + # Analysis # Analysis commands from srvyr are also similar to ones that work on local data.frames.