Subject Performance optimation?
Author checkmail



I save values in some tables (simpler description)


First a Table who saved the timestamp of the mensuration

Table A timestamps

ID primary key

TS timestamp


Second a Table with the measured data (25 records/measured sensors will be saved every 10 Minutes, one record in Table A, 25 in Table B)

Table B mensuration

ID primary key

ID_counter integer of item to measure

ID_Timestamp foreign key of Table A

Value (double precision)


Now I would like to make an analysis. At the time, I do this:


for select cast(ts as date) as mz from tablea where ts >= “criteria from” and ts < “criteria to”

    group by mz)

    into :messzeit do


      f_messwert = null;

      MESSWERTE = '';

      for select a.id_counter, sum(a.value) from tableb a left join tablea b on a.id_timestamp =

      where cast(b.ts as date) = :messzeit

      group by a.id_counter

      into :i_zae, :f_messwert do


        if(f_messwert is null) then f_messwert = 0;

        MESSWERTE = MESSWERTE || cast(:i_zae as varchar(4)) || '=' || cast(:f_messwert as varchar(8)) || ';';






The Result is one returned record for each day (day, conter 1 = 123; counter 2 = 222;…)


It takes a long time but I must integrate the tablea on the second part of the statement. How can I optimize this in firebird?


Thank you