Re: [sqlite] New column in select will not mask column of the same name in having clause and sqlite won't warn

2015-01-04 Thread Dan Kennedy
On 01/05/2015 02:52 PM, Dan Kennedy wrote: On 01/05/2015 01:39 PM, Hick Gunter wrote: This is completely legal and well defined. HAVING is applied to the RESULT set of a SELECT. The select asks to count the "distinct kontrola" in each group of kvadrat and datum, the HAVING clause specifies

Re: [sqlite] New column in select will not mask column of the same name in having clause and sqlite won't warn

2015-01-04 Thread Dan Kennedy
On 01/05/2015 01:39 PM, Hick Gunter wrote: This is completely legal and well defined. HAVING is applied to the RESULT set of a SELECT. The select asks to count the "distinct kontrola" in each group of kvadrat and datum, the HAVING clause specifies returning only those records with pocet > 1.

Re: [sqlite] New column in select will not mask column of the same name in having clause and sqlite won't warn

2015-01-04 Thread Hick Gunter
This is completely legal and well defined. HAVING is applied to the RESULT set of a SELECT. The select asks to count the "distinct kontrola" in each group of kvadrat and datum, the HAVING clause specifies returning only those records with pocet > 1. If there were no pocet column in table b,

[sqlite] Missing db name with the SQLITE_ATTACH action callback

2015-01-04 Thread gwenn
Hello, I am trying to implement a cache of one connection metadata (databases, tables, columns, indexes). It seems possible to automatically update the cache by using an authorizer. But there is one problem with the SQLITE_ATTACH action: #define SQLITE_ATTACH 24 /* Filename