[
https://issues.apache.org/jira/browse/IMPALA-15090?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=18113421#comment-18113421
]
ASF subversion and git services commented on IMPALA-15090:
----------------------------------------------------------
Commit 2c011ceba3b160746c064d4aceccf5d8a558fa74 in impala's branch
refs/heads/master from Gokul Kolady
[ https://gitbox.apache.org/repos/asf?p=impala.git;h=2c011ceba ]
IMPALA-15090: Make Query Profile Redaction DOM to DOM
Remove profile JSON serialization from redaction and
GenerateAiAnalysisFromProfile by extracting redaction inputs
from DOM strings and estimating profile size via DOM traversal.
Cache redacted profile size in QueryProfileRedactor and pass
it to profile parsing to avoid extra JsonToString() allocations
in the AI profile-analysis flow.
Testing:
Added targeted unit coverage in query-profile-redaction-test
for the DOM-based helper path: nested JSON string collection,
regex matching/dedup across collected DOM strings, IPv6
candidate filtering with validation, and serialized-size
estimation parity checks. Existing redaction regression tests
continue to validate SQL/user/IP/table/column/host masking and
unredaction behavior after removing profile text round-trips
from the analysis flow. Also validated the end-to-end AI
profile analysis feature on a local Impala setup.
Assisted-by: Codex 5.3 (Cursor)
Change-Id: I8b8ad721a5531e9d257cf6e54b83ea3ee9734039
Reviewed-on: http://gerrit.cloudera.org:8080/24429
Reviewed-by: Impala Public Jenkins <[email protected]>
Tested-by: Impala Public Jenkins <[email protected]>
> Make Query Profile Redaction DOM to DOM
> ---------------------------------------
>
> Key: IMPALA-15090
> URL: https://issues.apache.org/jira/browse/IMPALA-15090
> Project: IMPALA
> Issue Type: Improvement
> Reporter: Gokul Kolady
> Assignee: Gokul Kolady
> Priority: Major
>
> Currently, the query profile redaction module (used by the Impala AI Query
> Profile Analyzer) takes in a JSON DOM, converts to text, redacts the text,
> then converts back to JSON DOM. We should remove these conversions and
> instead maintain the DOM structure throughout redaction.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
