[SOLR-1498] RegexTransformer: sourceColName version not handling multiValued fields correctly

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.4
• Fix Version/s: 1.4
• Component/s: contrib - DataImportHandler

Environment

Windows XP, JDK 6, Tomcat 6
Linux (RedHat), JDK, Tomcat 5

• Assignee: Noble Paul
• Reporter: Chantal Ackermann
• Created: 08/Oct/09 14:43
• Updated: 10/Nov/09 15:52
• Resolved: 10/Oct/09 06:14

Description

Versions in use/compared:
Solr 1.3
(Nightly 5th August)
Nightly 22nd September

As RegexTransformer is not different between the two nightlies, the
issue probably appeared before.

ISSUE:
Using RegexTransformer with the 'sourceColName' notation will not populate
multiValued (actually containing multiple values) fields with a list but
instead add only one value per document.

The version with 'groupNames' does.

worked for 1.3 (regression):
<field column="participant" sourceColName="person" regex="([^\|]+)|.*" />
<field column="role" sourceColName="person"
regex="[^\|]|\d,\d+,\d+,(.*)" />

works for nightly 22nd Sept:
<field column="person" groupNames="participant,role"
regex="([^\|])|\d,\d+,\d+,(.*)" />

(Both fields are of type solr.StrField and multiValued.)

Comparing the source code of RegexTransformer 1.3 vs. 22nd Sept, I found:

for (Object result : results)
row.put(col, result);

(lines 106-107 of transformRow() 22nd of Sept)