[CASSANDRA-7095] Cqlsh reads floats and doubles with added non-trivial precision

• Type: Bug
• Status: Resolved
• Priority: Normal
• Resolution: Not A Problem
• Component/s: Legacy/Tools

Environment

Ubuntu 12.04.3
Python 2.7.3

• Severity: Normal
• Assignee: Mikhail Stepura
• Reporter: Andrew Spitzer

Authors

Mikhail Stepura

• Created: 25/Apr/14 21:30
• Updated: 16/Apr/19 09:31
• Resolved: 25/Apr/14 22:15

Description

I found that on Cqlsh v. 4.1.1 and Cassandra 2.0.6.28 when I inserted value into float or double column, "select" returned a slightly different value

If cqlsh returns with only a low level precision, this does not show:

cqlsh> create KEYSPACE test WITH replication = {'class': 'SimpleStrategy', 'replication_factor':1};
cqlsh> use test ;
cqlsh:test> create table testing (a int PRIMARY KEY , b float);
cqlsh:test> INSERT INTO testing (a, b ) VALUES (  1, 2.0002);
cqlsh:test> select * FROM testing ;

 a | b
---+--------
 1 | 2.0002

But if precision is increased, for example, 20 digits:

cqlsh> SELECT * FROM test.testing ;

 a | b
---+-----------------------
 1 | 2.0002000331878662109

Using sstable2json, one sees that this issue is with cqlsh, the actual data is the value inserted:

sstable2json /var/lib/cassandra/data/test/testing/test-testing-jb-1-Data.db 
[
{"key": "00000001","columns": [["","",1398459775180000], ["b","2.0002",1398459775180000]]}
]