You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The common theme across all table functions is that they can compute multiple output values for each input row. But regexp_extract seems to be a really special table function in how it returns its multiple output values. Most table functions put their multiple output values in multiple output rows. However, regexp_extract does the following:
When used in the FROM clause, it adds multiple columns.
When used in the SELECT clause, it seems to be creating a record (in one column).
I'd improve the docs myself, but I'm not sure about various things:
Is regexp_extract the only table function with the above properties?
Should we even consider it a table function if it's so different from all other table functions, or create a new category for it?
Do other databases have something similar? Note that Cockroach and Databricks have a function with the same name, but it only returns one output value, so it doesn't have the above peculiarities.
The text was updated successfully, but these errors were encountered:
The common theme across all table functions is that they can compute multiple output values for each input row. But
regexp_extract
seems to be a really special table function in how it returns its multiple output values. Most table functions put their multiple output values in multiple output rows. However,regexp_extract
does the following:FROM
clause, it adds multiple columns.SELECT
clause, it seems to be creating a record (in one column).I'd improve the docs myself, but I'm not sure about various things:
regexp_extract
the only table function with the above properties?The text was updated successfully, but these errors were encountered: