Details
-
Enhancement
-
Resolution: Done
-
Major
-
2.0.1.Final, 2.1.0.Alpha2
-
False
-
None
-
False
Description
A user reported that when he has this table
CREATE TABLE ks.tb ( s text, sp int, d text, dp int, t timestamp, m map<text, text>, PRIMARY KEY (s, sp, d, dp) ) WITH CLUSTERING ORDER BY (sp ASC, d ASC, dp ASC)
and he does this CQL:
DELETE FROM ks.tb WHERE s = '1,33619828888' and sp = 1 and d = '12, 23, toto' and dp > 1;
It will result in this record:
"after": { "s": { "value": "1,33619828888", "deletion_ts": null, "set": true }, "sp": null, "d": null, "dp": null, "m": null, "t": null, "_range_start": "EXCL_START_BOUND(1, 12, 23, toto, 1)", "_range_end": "INCL_END_BOUND(1, 12, 23, toto)" }
This is quite unfortunate to parse on consumer level because consumer does not know what to actually parse as values because there are commas in value and commas are separating the values themselves.
The solution is to do this:
"after": { "s": { "value": "1,33619828888", "deletion_ts": null, "set": true }, "sp": null, "d": null, "dp": null, "m": null, "t": null, "_range_start": { "method": "EXCL_START_BOUND", "values": { "sp": "1", "d": "12, 23, toto", "dp": "1" } }, "_range_end": { "method": "INCL_END_BOUND", "values": { "sp": "1", "d": "12, 23, toto" } } }
