[
https://issues.apache.org/jira/browse/PHOENIX-3560?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15813531#comment-15813531
]
Mujtaba Chohan commented on PHOENIX-3560:
-----------------------------------------
[~samarthjain] Patch didn't make much difference. Count * query remains 3X
slower for immutable encoded table. In my current test I used 10 byte random
column names for each of the 5K columns so data size for non-encoded table
wasn't comparatively smaller as in the original test where we used column names
as column_1 ... column_5000.
Couple of more data points. Data was loaded from OS page cache and I forced the
query to execute serially for comparison. Created those table with 20K rows and
each of those 5K columns were filled 15 bytes. I'll also attach data generator
script.
> Aggregate query performance is worse with encoded columns for schema with
> large number of columns
> -------------------------------------------------------------------------------------------------
>
> Key: PHOENIX-3560
> URL: https://issues.apache.org/jira/browse/PHOENIX-3560
> Project: Phoenix
> Issue Type: Sub-task
> Reporter: Mujtaba Chohan
> Assignee: Thomas D'Silva
> Fix For: 4.10.0
>
> Attachments: PHOENIX-3565.patch
>
>
> Schema with 5K columns
> {noformat}
> create table (k1 integer, k2 integer, c1 varchar ... c5000 varchar CONSTRAINT
> PK PRIMARY KEY (K1, K2))
> VERSIONS=1, MULTI_TENANT=true, IMMUTABLE_ROWS=true
> {noformat}
> In this test, there are no null columns and each column contains 200 chars
> i.e. 1MB of data per row.
> Count * aggregation is about 5X slower with encoded columns when compared to
> table non-encoded columns using the same schema.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
