[jira] [Commented] (IMPALA-14680) Improve row regex search syntax of runtime profile tests
[ https://issues.apache.org/jira/browse/IMPALA-14680?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18054782#comment-18054782 ] Dániel Gábor Vankó commented on IMPALA-14680: - Hi [~surya-hebbar] Have you thought about enhancing the result verifiers similarly as well? Currently 'row_regex:' works differently in the RUNTIME_PROFILE and in the RESULTS section. (The latter still uses regex.match through the [ResultRow's __eq__|https://github.com/apache/impala/blob/3dac0135fba0717dd977043e7ecc6b52bf55189f/tests/common/test_result_verifier.py#L165] function.) Or that would not be that much performance gain? > Improve row regex search syntax of runtime profile tests > > > Key: IMPALA-14680 > URL: https://issues.apache.org/jira/browse/IMPALA-14680 > Project: IMPALA > Issue Type: Improvement >Reporter: Surya Hebbar >Assignee: Surya Hebbar >Priority: Major > Fix For: Impala 5.0.0 > > > While enabling the aggregated profile IMPALA-9846, we improve the regex > search implementation for the runtime profile tests. > To improve this, we also need to improve the current `row_regex` search > syntax used for search. > This has greedy search syntax like ".*" at the beginning and end of > row_regex, this is unnecessary and needs to be removed. -- This message was sent by Atlassian Jira (v8.20.10#820010) - To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
[jira] [Commented] (IMPALA-14680) Improve row regex search syntax of runtime profile tests
[ https://issues.apache.org/jira/browse/IMPALA-14680?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18054409#comment-18054409 ] ASF subversion and git services commented on IMPALA-14680: -- Commit 00c233cc4fc25d23fc8a7e2f1efdf2d85c29f653 in impala's branch refs/heads/master from Daniel Vanko [ https://gitbox.apache.org/repos/asf?p=impala.git;h=00c233cc4 ] IMPALA-14692: Fix test_spilling_hash_join IMPALA-14680 mistakenly removed greedy regex patterns from query-impala-13138.test, but this test checks for the query results not for the query profile, which was modified in IMPALA-14680. Testing: * test_spilling_hash_join passed in exhaustive mode Change-Id: I709f81217f44c9377e4a1e8419787591ba7b7451 Reviewed-on: http://gerrit.cloudera.org:8080/23898 Reviewed-by: Impala Public Jenkins Tested-by: Impala Public Jenkins > Improve row regex search syntax of runtime profile tests > > > Key: IMPALA-14680 > URL: https://issues.apache.org/jira/browse/IMPALA-14680 > Project: IMPALA > Issue Type: Improvement >Reporter: Surya Hebbar >Assignee: Surya Hebbar >Priority: Major > Fix For: Impala 5.0.0 > > > While enabling the aggregated profile IMPALA-9846, we improve the regex > search implementation for the runtime profile tests. > To improve this, we also need to improve the current `row_regex` search > syntax used for search. > This has greedy search syntax like ".*" at the beginning and end of > row_regex, this is unnecessary and needs to be removed. -- This message was sent by Atlassian Jira (v8.20.10#820010) - To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
[jira] [Commented] (IMPALA-14680) Improve row regex search syntax of runtime profile tests
[ https://issues.apache.org/jira/browse/IMPALA-14680?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18052302#comment-18052302 ] ASF subversion and git services commented on IMPALA-14680: -- Commit 97d766577df69b5e602c811063f365e464390e21 in impala's branch refs/heads/master from Surya Hebbar [ https://gitbox.apache.org/repos/asf?p=impala.git;h=97d766577 ] IMPALA-14680: Improve row regex search syntax in runtime profile tests Currently, the runtime profile tests contain row regex searches which try to find matches by comparing the regex line by line. This form of search is inefficient. So, while updating the tests for the aggregated profile IMPALA-9846, this performance is being improved by accumulating row regexes together, then searching the entire profile at once. In order to support this improvement, we need to correct the current `row_regex` syntax being used. The current tests use greedy regex like ".*" at the beginning and end of `row_regex` searches. Using greedy regex in this way consumes more resources and is redundant for the current implementation. To fix this, these additional greedy regex characters(i.e. `.*`,`.+`) are being removed or replaced across all the runtime profile tests. Change-Id: I1460c2d22b03c06aa43c85f78fa9e05cec2775ec Reviewed-on: http://gerrit.cloudera.org:8080/23864 Tested-by: Impala Public Jenkins Reviewed-by: Csaba Ringhofer > Improve row regex search syntax of runtime profile tests > > > Key: IMPALA-14680 > URL: https://issues.apache.org/jira/browse/IMPALA-14680 > Project: IMPALA > Issue Type: Improvement >Reporter: Surya Hebbar >Assignee: Surya Hebbar >Priority: Major > > While enabling the aggregated profile IMPALA-9846, we improve the regex > search implementation for the runtime profile tests. > To improve this, we also need to improve the current `row_regex` search > syntax used for search. > This has greedy search syntax like ".*" at the beginning and end of > row_regex, this is unnecessary and needs to be removed. -- This message was sent by Atlassian Jira (v8.20.10#820010) - To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
