assign unique number to each row returned by regexp_split_to_table in Postgres

Question

I need below output for the regexp_split_to_table output: -

select  regexp_split_to_table ('a,b,c,d,e,f',',') expr
---------------------------------------------

rownum    expr
--------------
1          a
2          b
3          c
4          d
5          e
6          f

i tried using the row_number() function as well, but it is returning 1 for all the rows

select row_number() over (), regexp_split_to_table ('a,b,c,d,e,f',',') 

rownum    expr
--------------
1          a
1          b
1          c
1          d
1          e
1          f

Answer 1

That's what with ordinality is for:

select rownum, expr
from regexp_split_to_table ('a,b,c,d,e,f',',') with ordinality as t(expr, rownum)

The rownum ("ordinality") is the "index" of that element in the result of regexp_split_to_table

---

Note that using string_to_array() with unnest is typically faster as regexes are quite expensive:

select rownum, expr
from unnest(string_to_array('a,b,c,d,e,f',',')) with ordinality as t(expr, rownum)

Answer 2

Try wrapping your current query in a CTE and then use ROW_NUMBER :

WITH cte AS (
    SELECT REGEXP_SPLIT_TO_TABLE ('a,b,c,d,e,f', ',') expr
)

SELECT
    ROW_NUMBER() OVER (ORDER BY expr) rownum,
    expr
FROM cte
ORDER BY expr;