Details
-
New Feature
-
Status: Open
-
Minor
-
Resolution: Unresolved
-
3.5.0
-
None
-
None
Description
SparkSQL supports regex_column_names.
https://spark.apache.org/docs/latest/sql-ref-syntax-qry-select.html
However, support depends on a config: spark.sql.parser.quotedRegexColumnNames
The reason is that it overloads proper identifier names.
Here we propose a cleaner, compatible API:
SELECT * LIKE 'pattern' ...
The semantic should follow common regular expression patterns used for the LIKE operator with the caveat that it should obey identifier case insensitivity setting.