Google BigQuery: iterate CONTAINS function over subquery - sql

Question

Lets assume I have two tables:
girls prefixes
------ ----------
Le-na -na
Lo-ve -ve
Li-na -la
Lu-na -ta
Len-ka -ya
All girls names and prefixes are different length!
I want to select all girl names that contains prefixes table and to do it in a query(imagine I have many names and many prefixes).
I untested that for single case it is being completed like this:
SELECT girls,SOME(girls CONTAINS ("-na")) WITHIN RECORD FROM prefixes
But how do I implement iteration of CONTAINS function over subquery?
e.g.
SELECT girls,SOME(girls CONTAINS (SELECT * FROM prefixes))
WITHIN RECORD FROM prefixes
–– this doesn't work, cause Subselect not allowed in SELECT clause
I'd really appreciate any ideas, I've tried to search for this but couldn't find my case.

Answer 1

Have you tried just using join?
select *
from girls g join
prefixes p
on g.girls like concat('%', p.prefix);
This should work using standard SQL.

Answer 2

Assuming that the prefixes (well, suffixes) are always three characters, you can perform an efficient semi-join with the result of SUBSTR:
#standardSQL
WITH Girls AS (
SELECT name
FROM UNNEST(['Le-na', 'Lo-ve', 'Li-na', 'Lu-na', 'Len-ka']) AS name
),
Suffixes AS (
SELECT suffix
FROM UNNEST(['-na', '-ve', '-la', '-ta', '-ya']) AS suffix
)
SELECT
name
FROM Girls
WHERE EXISTS (
SELECT 1 FROM Suffixes WHERE suffix = SUBSTR(name, LENGTH(name) - 2)
);
Or you can use LIKE, but it is equivalent to performing a cross join with a filter, so it probably won't be as fast:
#standardSQL
WITH Girls AS (
SELECT name
FROM UNNEST(['Le-na', 'Lo-ve', 'Li-na', 'Lu-na', 'Len-ka']) AS name
),
Suffixes AS (
SELECT suffix
FROM UNNEST(['-na', '-ve', '-la', '-ta', '-ya']) AS suffix
)
SELECT
name
FROM Girls
WHERE EXISTS (
SELECT 1 FROM Suffixes WHERE name LIKE CONCAT('%', suffix)
);
Edit: another option that enumerates all name suffixes for use in the semi-join:
#standardSQL
WITH Girls AS (
SELECT name
FROM UNNEST(['Le-na', 'Lo-ve-lala', 'Li-na', 'Lu-eya', 'Len-ka']) AS name
),
Suffixes AS (
SELECT suffix
FROM UNNEST(['-na', '-ve', '-lala', '-ta', '-eya']) AS suffix
),
GirlNamePermutations AS (
SELECT name, SUBSTR(name, LENGTH(name) + 1 - len) AS name_suffix
FROM Girls
CROSS JOIN UNNEST(GENERATE_ARRAY(1, (SELECT MAX(LENGTH(suffix)) FROM Suffixes))) AS len
)
SELECT
name
FROM GirlNamePermutations
WHERE EXISTS (
SELECT 1
FROM Suffixes
WHERE suffix = name_suffix
);
If you know the range of suffix lengths, you could hard-code it instead, e.g. replace:
CROSS JOIN UNNEST(GENERATE_ARRAY(1, (SELECT MAX(LENGTH(suffix)) FROM Suffixes))) AS len
with:
CROSS JOIN UNNEST(GENERATE_ARRAY(1, 5)) AS len

Answer 3

Below is for BigQuery Standard SQL
#standardSQL
WITH girls AS (
SELECT name
FROM UNNEST(['Le-na', 'Lo-ve', 'Li-na', 'Lu-na', 'Len-ka']) AS name
),
suffixes AS (
SELECT suffix
FROM UNNEST(['-na', '-ve', '-la', '-ta', '-ya']) AS suffix
)
SELECT name
FROM girls
JOIN suffixes
ON ENDS_WITH(name, suffix)
as an option - in case you will need to extend this to find fragments inside name - you can use REGEXP_CONTAINS
SELECT name
FROM girls
JOIN suffixes
ON REGEXP_CONTAINS(name, suffix)
or - STARTS_WITH to match by prefixes (vs. suffixes)
SELECT name
FROM girls
JOIN suffixes
ON STARTS_WITH(name, suffix)