[GRASS-dev] SQLite driver problem

[Moritz Lennert]

On 14/12/10 16:49, Markus Neteler wrote:
> On Tue, Dec 14, 2010 at 4:43 PM, Moritz Lennert
> <mlennert at club.worldonline.be>  wrote:
>> On 14/12/10 16:11, Markus Neteler wrote:
>>>
>>> Hi,
>>>
>>> I have some troubles with the SQLite driver with a SELECT statement:
>>>
>>> GRASS 6.4.1svn (patGB1):~>    echo 'SELECT cat FROM
>>> MRR1_09_marked_csv_lines_id ' | db.execute
>>
>>> I wonder what the problem could be...
>>
>> s/db.execute/db.select
>
> Cool.
>
> But I thought that both did essentially the same job? Apparently not...

No, db.select is only for selection and db.execute is for any other db 
manipulations (update, delete, insert, create table, etc). They use 
different functions to interact with the DB. Don't know what Radim's 
reasoning was behind this, but I guess he had a good one :-).

> I wonder
> if we could trap "SELECT" in db.execute and point the user to db.select.
> ?

That's not as easy as it sounds, since you might have SELECT clauses in 
UPDATE and other db.execute clauses (e.g. INSERT INTO mytable SELECT x, 
y, z FROM othertable). So the trapping would have to be a bit more 
sophisticated, at least looking out for a SELECT at the beginning of the 
statement. But definitely possible. Probably best done right before the

ret = db_execute_immediate(driver, &stmt);

at
line 74 of db/base/execute.c (grass6)
or
line 75 of /db/db.execute/main.c (grass_trunk)


Moritz