This adds support for the views subqueries patch at http://drupal.org/node/143888 (this is required for views_fastsearch) - note that this has nothing to do with views_union subqueries :)
The support is transparent - if the views subqueries patch is not present nothing will be done. I think this would be good to add because the kind of people who need complex things like views_union also likely need complex things like views_fastsearch!
| Comment | File | Size | Author |
|---|---|---|---|
| subquery.patch | 647 bytes | owen barton |
Comments
Comment #1
Anonymous (not verified) commentedI'm hesitant about this. This would cause unexpected problems if the phrase [SUBQUERY] appears in the data. It seems like we'd need some more sophisticated parsing, so that it would only change the actual subqueries and not mentions of [SUBQUERY] in the data.
Let me see...
A query that it works on would be like this?
SELECT thing FROM {a} A, [SUBQUERY] B WHERE A.blah=B.thing AND A.blah='Hello, World';
right?
So something like this would still work:
SELECT thing FROM {a} A, [SUBQUERY] B WHERE A.blah=B.thing AND A.blah='I\'ve got a cool [SUBQUERY]';
because the str_replace would only replace the first instance of [SUBQUERY] and it would be correct.
But something like:
SELECT thing, 'check out my [SUBQUERY]' FROM {a} A, [SUBQUERY] B WHERE A.blah=B.thing AND A.blah='Hello, World';
is valid SQL but this patch wouldn't do it correctly.
It may be the case that we could never produce a query like that, and that all the data would occur after all possible instances of [SUBQUERY]. Can we prove that?