[DRILL-1645] Can not query any column if the row size is larger than 128K

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.6.0
• Fix Version/s: 0.7.0
• Component/s: Query Planning & Optimization
• Assignee: Unassigned
• Reporter: Hao Zhu
• Created: 06/Nov/14 17:27
• Updated: 04/Jan/15 23:58
• Resolved: 04/Jan/15 23:57

Description

If the row size is larger than 128K, drill can not query any column no matter how small the column is.
So once the top level row of the json file is larger than 128K, then we can not query anything from this file.

We need drill can query small columns which is below the limitation at least.

Reproduce:

0: jdbc:drill:> select * from `wrapper.json`;
Query failed: Failure while running fragment. Record is too long. Max allowed record size is 131072 bytes. [7894614e-9928-4aa5-b649-2d085d01a55b]

Error: exception while executing query: Failure while trying to get next result batch. (state=,code=0)
0: jdbc:drill:> select id from `wrapper.json`;
Query failed: Failure while running fragment. Record is too long. Max allowed record size is 131072 bytes. [ed61693e-d4da-478e-b85d-1fb7d9621243]

Error: exception while executing query: Failure while trying to get next result batch. (state=,code=0)

The sample data format is like below(Complete sample data is attached.):
In above test case, the "id" column should be fetched since it is very small.

[root@maprdemo data]# more wrapper.json 
{"id": 1,"supercolumn": {"content":testtesttesttesttesttestt......."}}

Issue Links

• relates to: DRILL-1632 Increase Maximum Allowed Record Size From 131072 bytes