Am 22.09.2017 um 10:43 schrieb Aki Tuomi:
> To my inexpert eye it seems that in your original post, 4.1 was fastest
> of the lot, so the database begin/commit + prepare cannot really explain
> this since this has not been removed in 4.1?
Obviously there were some changes - no more BEGIN/COMMIT
Update: I do see that the BEGIN/COMMIT exists in 4.0, but was removed in
Master (4.1pre). That might explain why Master is faster :-)
regards
Klaus
Am 22.09.2017 um 10:35 schrieb Klaus Darilion:
>
>
> Am 21.09.2017 um 20:20 schrieb Aki Tuomi:
>> Statements are supposed to prepared once, not per
On 22.09.2017 11:35, Klaus Darilion wrote:
>
> Am 21.09.2017 um 20:20 schrieb Aki Tuomi:
>> Statements are supposed to prepared once, not per every query.
> Indeed, that is the case, the "prepare" is only done once.
>
> But I see that simple SELECTs are wrapped into transactions. Maybe this
> is
Am 21.09.2017 um 20:20 schrieb Aki Tuomi:
> Statements are supposed to prepared once, not per every query.
Indeed, that is the case, the "prepare" is only done once.
But I see that simple SELECTs are wrapped into transactions. Maybe this
is also a performance penalty:
regards
Klaus
LOG: dur
