Don’t we all command line workers get angry when a command google:”can’t read from stdin”? I thought oracle’s google:”tkprof” is one such command. I was wrong! This is very useful when there are many small SQL trace files or when one wants to tkprof zipped trace file as is, like below. I checked the generated file content briefly and as far as I can tell, aggregation was done correctly.
$ echo *.trc | wc -w 220 $ cat *.trc | wc -l 3391365 $ cat *.trc | tkprof /dev/stdin all.tkprof $ tail all.tkprof 280 sessions in tracefile. 3150861 user SQL statements in trace file. 79974 internal SQL statements in trace file. 17923 SQL statements in trace file. 4434 unique SQL statements in trace file. 3391365 lines in trace file.
$ (gzcat orcl_ora_9120.trc.gz && gzcat orcl_ora_9845.trc.gz) | wc -l 6829778 $ (gzcat orcl_ora_9120.trc.gz && gzcat orcl_ora_9845.trc.gz) | tkprof /dev/stdin /tmp/test.tkprof $ tail /tmp/test.tkprof 3 sessions in tracefile. 50078 user SQL statements in trace file. 47 internal SQL statements in trace file. 30583 SQL statements in trace file. 8920 unique SQL statements in trace file. 6829778 lines in trace file. 15995 elapsed seconds in trace file.
Advertisement
