Обсуждение: libpq: multiple commands within single query
By which rules database returns results for multiple commands within single query?
For example I send (execute) such query (obtain different information about just connected client):
"SELECT column1 FROM table1; SELECT column2 FROM table2; SELECT column3 FROM table3;"
And when I get response I expect some results (PGresult structure) for EACH "SELECT" - so there must be 3 PGresult structures and if one of results contains zero rows - it means that there are no data (of such type) for such client yet, but there still can be some data for following "SELECT". For example (referring to previous example) response can contain 3 results (as I expect) where first contains 1 row, second 0 rows and third 1 row. So when I examine the response I can match each PGresult to each "SELECT".
Does it works on this way? Or I missed something?
For example I send (execute) such query (obtain different information about just connected client):
"SELECT column1 FROM table1; SELECT column2 FROM table2; SELECT column3 FROM table3;"
And when I get response I expect some results (PGresult structure) for EACH "SELECT" - so there must be 3 PGresult structures and if one of results contains zero rows - it means that there are no data (of such type) for such client yet, but there still can be some data for following "SELECT". For example (referring to previous example) response can contain 3 results (as I expect) where first contains 1 row, second 0 rows and third 1 row. So when I examine the response I can match each PGresult to each "SELECT".
Does it works on this way? Or I missed something?
Hey,
--
// Dmitriy.
18 января 2011 г. 14:24 пользователь Вячеслав Блинников <slavmfm@gmail.com> написал:
By which rules database returns results for multiple commands within single query?
For example I send (execute) such query (obtain different information about just connected client):
"SELECT column1 FROM table1; SELECT column2 FROM table2; SELECT column3 FROM table3;"
And when I get response I expect some results (PGresult structure) for EACH "SELECT" - so there must be 3 PGresult structures and if one of results contains zero rows - it means that there are no data (of such type) for such client yet, but there still can be some data for following "SELECT". For example (referring to previous example) response can contain 3 results (as I expect) where first contains 1 row, second 0 rows and third 1 row. So when I examine the response I can match each PGresult to each "SELECT".
Does it works on this way? Or I missed something?
According the documentation: "the client can be handling the
results of one command while the server is still working on later
queries in the same command string.". The key word here is "later".
results of one command while the server is still working on later
queries in the same command string.". The key word here is "later".
--
// Dmitriy.
Yeah - that is how I do (PGgetResult function returns results until it will return "NULL" pointer which means that all data for whole "PGsendQuery" was returned). But what does it mean when query which contain 3 "SELECT" commands returns just one result (second "PGgetResult()" already returns null pointer)?
18 января 2011 г. 14:55 пользователь Dmitriy Igrishin <dmitigr@gmail.com> написал:
Hey,18 января 2011 г. 14:24 пользователь Вячеслав Блинников <slavmfm@gmail.com> написал:By which rules database returns results for multiple commands within single query?
For example I send (execute) such query (obtain different information about just connected client):
"SELECT column1 FROM table1; SELECT column2 FROM table2; SELECT column3 FROM table3;"
And when I get response I expect some results (PGresult structure) for EACH "SELECT" - so there must be 3 PGresult structures and if one of results contains zero rows - it means that there are no data (of such type) for such client yet, but there still can be some data for following "SELECT". For example (referring to previous example) response can contain 3 results (as I expect) where first contains 1 row, second 0 rows and third 1 row. So when I examine the response I can match each PGresult to each "SELECT".
Does it works on this way? Or I missed something?According the documentation: "the client can be handling the
results of one command while the server is still working on later
queries in the same command string.". The key word here is "later".
--
// Dmitriy.
=?KOI8-R?B?99HexdPMwdcg4szJzs7Jy8/X?= <slavmfm@gmail.com> writes:
> Yeah - that is how I do (PGgetResult function returns results until it will
> return "NULL" pointer which means that all data for whole "PGsendQuery" was
> returned). But what does it mean when query which contain 3 "SELECT"
> commands returns just one result (second "PGgetResult()" already returns
> null pointer)?
Most likely, the command failed ... have you tried inspecting the
PGresult to see what status it's reporting?
regards, tom lane