"Guillaume Smet" <[EMAIL PROTECTED]> writes: > As explained in the previous thread, we used to play with log_duration > and log_min_error_statement to have the following behaviour: > - log every duration so that we can have a global overview of the > database activity; > - log statement only for slowest queries.
I really find it pretty bizarre to want to log a duration without logging the statement that caused it. Seems like the log_min_duration_statement parameter ought to be enough. These logging options are already messy and unintuitive, and adding still more odd frammishes doesn't help that ... Changing code without changing the associated comments isn't a good way to get your ideas accepted, either, eg here: /* Only print duration if we previously printed the statement. */ ! if ((log_statement == LOGSTMT_ALL && save_log_duration == LOGDRTN_LOGGED) || save_log_duration == LOGDRTN_ALL) ereport(LOG, The patch makes that comment a lie. This is not acceptable coding practice. regards, tom lane ---------------------------(end of broadcast)--------------------------- TIP 9: In versions below 8.0, the planner will ignore your desire to choose an index scan if your joining column's datatypes do not match