[CALCITE-5580] Add SPLIT function (enabled in BigQuery library)

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.35.0
• Labels: pull-request-available
• Assignee: Tanner Clary
• Reporter: Tanner Clary
• Created: 14/Mar/23 02:32
• Updated: 01/Aug/23 07:29
• Resolved: 22/Mar/23 12:57

Description

BigQuery offers the SPLIT() function which splits a string at an optionally-specified delimiter into a string array. If no delimiter is specified, it is default to a comma. If the string is empty, an array of a single empty string is returned. If the delimiter is not found in the string, an array with a single element (the string) is returned.

In BigQuery, the function can also accept bytes. In order to implement this, I think some modifications to ByteString.java may be required. I will probably not do this at least for my initial draft. If anyone has any suggestions or guidance on whether or not it should be supported, I would appreciate it.

Documentation and example cases may be found below.

EXAMPLE: SPLIT('h,e,l,l,o') would return: [h, e, l, l, o].
EXAMPLE: SPLIT('h-e-l-l-o', '-') would return: [h, e, l, l, o].

BigQuery docs

Issue Links

• relates to: CALCITE-5879 AssertionError during constant reduction of SPLIT expression that returns NULL
• relates to: CALCITE-5882 Compile-time evaluation of SPLIT function returns incorrect result
• links to: GitHub Pull Request #3109