Hello all,

Here a simple issue I have with performance :

1/ I create a table A and populate it with a bulk load

2/I create a table B, I populate the table with bulk load

3/I create table C with a CTAS statement. I analyze the statistics and set the primary key

4/I try a “count distinct” type query (on all the table) on the PK Field. It is pretty slow, with a high sollicitation of my hard drive for several dozens of second.
the “select count” is however fast.

So I have these following questions :
1/why despite statistics the count distinct query is slow? I thought the idea of statistics was to get some metrics about the table?
2/if the field I make a count distinct on is the PK (or the set of field is the PK), it’s exactly the same than a count non null. So why the time are not the same?
3/if these issues are design issue or bugs, how can I ask to make it solved?

For information : Monetdb 11.31.13 on Windows 10

Best regards,

Simon AUBERT

Bourse Maritime – 1, place Lainé – 33 000 Bordeaux

Mob : +33(0)6 66 28 52 04