[CASSANDRA-15679] cqlsh COPY FROM of map of blobs fails with parse error "unhashable type: 'bytearray'" - ASF JIRA

XML

Word

Printable

JSON

Details

Type: Bug
Status: Resolved
Priority: Normal
Resolution: Fixed
Fix Version/s: 2.1.x, 2.2.17, 3.11.7, 4.0-alpha4, 4.0
Component/s: Tool/cqlsh
Labels:
- pull-request-available

Bug Category:
Code - Bug - Unclear Impact
Severity:
Low
Complexity:
Normal
Discovered By:
User Report
Platform:

All
Impacts:

None
Since Version:

2.1.21
Source Control Link:

https://gitbox.apache.org/repos/asf?p=cassandra.git;a=commit;h=f3568c0d50ac7573b53f0043b2567bde3b39bee8
Test and Documentation Plan:

Hide

This bug is triggered when we try to deserialize a blob inside a container, such as list, set, or map. copyutil stores blobs as bytearrays. Python bytearrays are mutable and not hashable data structures.

In Python 3 there is bytes type that is exactly what we need in this situation. On older versions that do not support Python 3 and in case we run cqlsh with Python 2 at 4.0, we should fall back to str.

https://github.com/apache/cassandra/pull/506
https://github.com/apache/cassandra/pull/507
https://github.com/apache/cassandra/pull/508
https://github.com/apache/cassandra/pull/509
https://github.com/apache/cassandra/pull/510

Show
This bug is triggered when we try to deserialize a blob inside a container, such as list, set, or map. copyutil stores blobs as bytearrays. Python bytearrays are mutable and not hashable data structures. In Python 3 there is bytes type that is exactly what we need in this situation. On older versions that do not support Python 3 and in case we run cqlsh with Python 2 at 4.0, we should fall back to str. https://github.com/apache/cassandra/pull/506 https://github.com/apache/cassandra/pull/507 https://github.com/apache/cassandra/pull/508 https://github.com/apache/cassandra/pull/509 https://github.com/apache/cassandra/pull/510

Description

Background

A user was having issues loading CSV data with the COPY FROM command into a map column with blob values.

Replication steps

I can easily replicate the problem with this simple table:

CREATE TABLE community.blobmaptable (
    id text PRIMARY KEY,
    blobmapcol map<int, blob>
)

I have this CSV file that contains just 1 row:

$ cat blobmap.csv 
c3,{3: 0x74776f}

And here's the error when I try to load it:

cqlsh:community> COPY blobmaptable (id, blobmapcol) FROM '~/blobmap.csv' ;
Using 1 child processes
Starting copy of community.blobmaptable with columns [id, blobmapcol].
Failed to import 1 rows: ParseError - Failed to parse {3: 0x74776f} : unhashable type: 'bytearray',  given up without retries
Failed to process 1 rows; failed rows written to import_community_blobmaptable.err
Processed: 1 rows; Rate:       2 rows/s; Avg. rate:       3 rows/s
1 rows imported from 1 files in 0.389 seconds (0 skipped).

I've also logged PYTHON-1234 because I wasn't sure if it was a Python driver issue. Cheers!