结构和数组 - 错误：标量子查询产生多个元素

Question

I'm trying to put two tables together using struct and array. My idea is for each row in table A apply Levenshtein Distance to table B.

Table A:

col1
whisky
delta
Tango

Table B

col1
Whiskey
delta force
Tango is great

Desired output:

col1        col2             col3
whisky    Whiskey            <lv_distance_score>
          delta force        <lv_distance_score>
          Tango is great     <lv_distance_score>
delta     Whiskey            <lv_distance_score>
          delta force        <lv_distance_score>
          Tango is great     <lv_distance_score>
Tango     Whiskey            <lv_distance_score>
          delta force        <lv_distance_score>
          Tango is great     <lv_distance_score>

For this, first I'm trying to just get de desired output of col1 and col2, but I keep getting an error that says Scalar subquery produced more than one element .

The query I wrote is:

WITH a AS (
SELECT col1, [STRUCT((SELECT col1 FROM table_B))] AS col2 FROM table_A
)
SELECT col1,c2 FROM a,UNNEST(a.col2) AS c2;

What I'm doing wrong here? How can I achieve what I'm looking for?

Answer 1

I'm a bit lost. Why not just use a cross join ?

select a.col1, b.col1
from a cross join
     b

If you want one row per row in a with an array for b, then:

select a.col1, array_agg(b)
from a cross join
     b
group by a.col1;

Answer 2

What I'm doing wrong here?

Below is simple fix for your original query

WITH a AS (
SELECT col1, 
  [STRUCT(ARRAY(SELECT col1 FROM table_B) as col2)] AS col2 
FROM table_A
)
SELECT col1, c2.col2 
FROM a, UNNEST(a.col2) AS c2;   

While above hopefully shows you what was wrong with your query - I am not sure it is right direction to go.

How can I achieve what I'm looking for?

you just go with simple cross join like in below example

SELECT a.col1, ARRAY_AGG(b.col1 ORDER BY lv_distance_score(a.col1, b.col1) LIMIT 1)
FROM table_A a 
CROSS JOIN table_B b
GROUP BY a.col1   

Note: you can find plenty of examples for Levenshtein Distance UDF here on SO