I've used Excel to generate numerous SELECT
statements from a list of the schema names from a database with a large number of identical schemas:
select result from foo.table limit 1;
select result from bar.table limit 1;
select result from doo.table limit 1;
(foo
, bar
& doo
are examples of my schemas, there are hundreds in reality).
Each SELECT
will return only one result. I simply want one column result
with as many rows as there are schemas. I can then copy this back into Excel against the schema names.
When I run the query above I get 1 row, with the others being discarded:
Query result with 1 row discarded. Query result with 1 row discarded. Total query runtime: 40 ms. 1 row retrieved.
I have tried using UNION ALL
, but the limit 1
I am using to ensure one row only is returned from each schema table appears to prevent this from working.
How can I either prevent the other rows from being discarded, or write a query that will return the values I need (two columns - schema_name, result - one row for each schema) in a more efficient way?
See Question&Answers more detail:os