You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@hive.apache.org by "Krisztian Kasa (Jira)" <ji...@apache.org> on 2023/05/17 13:28:00 UTC
[jira] [Created] (HIVE-27353) Show saved snapshot of materialized view source tables
Krisztian Kasa created HIVE-27353:
-------------------------------------
Summary: Show saved snapshot of materialized view source tables
Key: HIVE-27353
URL: https://issues.apache.org/jira/browse/HIVE-27353
Project: Hive
Issue Type: Improvement
Components: Materialized views
Reporter: Krisztian Kasa
Assignee: Krisztian Kasa
HIVE-25745 introduced a new section into
{code:java}
DESCRIBE FORMATTED <view name>;
{code}
output:
{code:java}
# Materialized View Source table information
Table name I/U/D since last rebuild
hive.default.src_txn 0/0/0
hive.default.src_txn_2 0/0/0
{code}
Unfortunately transactional stats are not reliable because such stats are supposed to be saved along with basic stats.
If something blocks saving the stats like
{code:java}
set hive.stats.autogather=false;
{code}
basic stats still can be refreshed using
{code:java}
analyze table <table> compute statistics;
{code}
but it won't collect and update transactional since the amount of rows affected by recent transactions are no longer available and can not be calculated.
The goal of this jira is to print the saved snapshot information of each source table instead:
* writeId in case of native acid tables
* snapshotId in case of iceberg tables
--
This message was sent by Atlassian Jira
(v8.20.10#820010)