pg-validate

Given table [foo smallint] [bar varchar(10)]

var metadata = ... // use pg-metadata module
var validate = require('pg-validate')

// Validate for PostgreSQL
var errors = validate.pg.object({
    foo: 1280000,
    bar: 'aaaaaa'
}, metadata)

// Validate for Redshift
var errors = validate.redshift.object({
    foo: 1280000,
    bar: 'aaaaaa'
}, metadata)

// Or the verbose way
var errors = validate.object({
    foo: 1280000,
    bar: 'aaaaaa'
}, metadata, { platform: validate.REDSHIFT })

Supported Types

Integer

• smallint or int2
• integer or int4
• bigint or int8
• serial
• bigserial

Numeric

• numeric or decimal
• numeric(p, s), where p is the maximum precision in decimal digits and s is the scale (number of fractional digits). numeric(p) selects a scale of 0.
• float4 or real
• float8 or double_precision or float
• PostgreSQL only: float(p), where p is the minimum acceptable precision in binary digits. float(1) to float(23) selects real, float(24) to float(53) selects double_precision. Note that p is not used to validate values, unlike numeric(p).

Boolean

• boolean

Char

• varchar or char or bpchar
• text: strings of any length on PostgreSQL, an alias of varchar on Redshift

Timestamp

DateTime

• timestamp[ (p) ] or timestamptz [ (p) ]
• date
• time [ (p) ] or timetz [ (p) ]
• interval [ fields ] [ (p) ]

For the purposes of validation, there's no difference between time and timetz, or timestamp and timestamptz. This is because PostgreSQL silently ignores time zones for types without a time zone. And though Redshift stores dates in UTC, it does accept input values with a time zone - with the exception of full time zone names like "America/New_York".

Not currently supported:

• Julian Day ("J2451187"), BC/AC dates, or ambiguous dates in a DateStyle mode
• Time values with a precision greater than 3
• Dates outside the range of moment.js and Date
• Custom configured timezone names or abbreviations. In the future, the list of supported timezone values could be dynamically generated, via SELECT row_to_json(pg_timezone_abbrevs) FROM pg_timezone_abbrevs and the same for pg_timezone_names. For those timezones that are DST sensitive (is_dst: true), we should validate that values have a date.
• POSIX-style time zone specifications

Test

Requires a PostgreSQL or Redshift database. Create a dummy user and database (nothing is written by the tests) and specify the credentials in the PG_VALIDATE environment variable:

PG_VALIDATE=postgres://user:password@localhost:5432/database
npm test