Details
-
Improvement
-
Status: Open
-
Minor
-
Resolution: Unresolved
-
Impala 2.0
-
None
Description
Pretty printing is quite slow. For example, running a query from impala-shell with and w/o -B makes the query run in 14.5s and 2.5s respectively.
Below are part of the profile for the two runs
– run impala-shell (pretty-printing)
ExecSummary: Operator #Hosts Avg Time Max Time #Rows Est. #Rows Peak Mem Est. Peak Mem Detail ------------------------------------------------------------------------------------------------------------------------- 05:EXCHANGE 1 1.509ms 1.509ms 6.37K 15 0 -1.00 B UNPARTITIONED 02:AGGREGATE 4 186.561ms 202.751ms 6.37K 15 14.26 MB 10.00 MB FINALIZE 04:AGGREGATE 4 8.469ms 9.46ms 6.38K 15 14.50 MB 10.00 MB 03:EXCHANGE 4 259.948us 345.133us 11.88K 15 0 0 HASH(score_card_full_parque... 01:AGGREGATE 4 172.837ms 185.338ms 11.88K 15 83.46 MB 10.00 MB 00:SCAN HDFS 4 355.90ms 499.316ms 56.62K 15 91.37 MB 432.00 MB default.score_card_full_par... Query Timeline: 14s823ms - Start execution: 108.674us (108.674us) - Planning finished: 16.688ms (16.579ms) - Submit for admission: 31.348ms (14.660ms) - Completed admission: 31.592ms (243.999us) - Ready to start remote fragments: 39.506ms (7.914ms) - Remote fragments started: 454.784ms (415.277ms) - Rows available: 2s031ms (1s576ms) - First row fetched: 2s065ms (33.861ms) - Unregister query: 14s817ms (12s751ms) ImpalaServer: - ***ClientFetchWaitTimer: 12s735ms*** <-- high - RowMaterializationTimer: 44.388ms
– run impala-shell -B (w/o pretty printing)
ExecSummary: Operator #Hosts Avg Time Max Time #Rows Est. #Rows Peak Mem Est. Peak Mem Detail ------------------------------------------------------------------------------------------------------------------------- 05:EXCHANGE 1 1.527ms 1.527ms 6.37K 15 0 -1.00 B UNPARTITIONED 02:AGGREGATE 4 180.55ms 194.285ms 6.37K 15 14.26 MB 10.00 MB FINALIZE 04:AGGREGATE 4 8.239ms 9.548ms 6.38K 15 14.52 MB 10.00 MB 03:EXCHANGE 4 292.4us 366.995us 11.88K 15 0 0 HASH(score_card_full_parque... 01:AGGREGATE 4 191.19ms 199.422ms 11.88K 15 83.46 MB 10.00 MB 00:SCAN HDFS 4 324.740ms 460.908ms 56.62K 15 91.37 MB 432.00 MB default.score_card_full_par... Query Timeline: 2s440ms - Start execution: 91.531us (91.531us) - Planning finished: 19.494ms (19.402ms) - Submit for admission: 32.238ms (12.743ms) - Completed admission: 32.377ms (139.598us) - Ready to start remote fragments: 40.827ms (8.449ms) - Remote fragments started: 457.854ms (417.27ms) - Rows available: 2s038ms (1s581ms) - First row fetched: 2s065ms (26.836ms) - Unregister query: 2s435ms (369.575ms) ImpalaServer: - ClientFetchWaitTimer: 342.945ms - RowMaterializationTimer: 47.580ms