[CASSANDRA-15905] cqlsh not able to fetch all rows when in batch mode

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Fixed
• Fix Version/s: 3.11.7, 4.0-beta1, 4.0
• Component/s: Legacy/CQL
• Bug Category: Degradation - Performance Bug/Regression
• Severity: Normal
• Complexity: Low Hanging Fruit
• Discovered By: User Report
• Platform: All
• Since Version: 3.10

Source Control Link

https://github.com/apache/cassandra/commit/9251b8116ff89b528b6b9eaa43d4dc2d1bc0bbaf

Test and Documentation Plan

dtest added

• Assignee: Yifan Cai
• Reporter: Yifan Cai

Authors

Yifan Cai

Reviewers

Brandon Williams

• Created: 26/Jun/20 00:56
• Updated: 03/Jan/21 17:02
• Resolved: 30/Jun/20 23:00

Description

The cqlsh in trunk only display the first page when running in the batch mode, i.e. using -execute or -file option. 
 
It is a change of behavior. In 3.x branches, the cqlsh returns all rows. 
 
It can be reproduced in 3 steps.

 1. ccm create trunk -v git:trunk -n1 && ccm start
 2. tools/bin/cassandra-stress write n=1k -schema keyspace="keyspace1"   // write 1000 rows
 3. bin/cqlsh -e "SELECT * FROM keyspace1.standard1;"                    // fetch all rows

 
There are 1000 rows written. But the output in step 3 will only list 100 rows, which is the first page. 

➜ bin/cqlsh -e "SELECT * FROM keyspace1.standard1" | wc -l
     105

 
The related change was introduced in https://issues.apache.org/jira/browse/CASSANDRA-11534, where the cqlsh.py script no longer fetch all rows when not using tty in the print_result method. 

Issue Links

• is related to: CASSANDRA-15537 4.0 quality testing: Local Read/Write Path: Upgrade and Diff Test