Balaji Manoharan created SOLR-5288:
--------------------------------------
Summary: Delta import is calling applyTranform() during
deltaQuerry and causing exception
Key: SOLR-5288
URL: https://issues.apache.org/jira/browse/SOLR-5288
Project: Solr
Issue Type: Bug
Components: contrib - DataImportHandler
Affects Versions: 4.4
Reporter: Balaji Manoharan
Priority: Critical
Fix For: 4.4
While experimenting delta import, was getting Script Exception such as
'toString()' is not found on null.
These are the queries that am using
a) Query > SELECT PK_FIELD, JOIN_DATE, USER_NAME FROM USERS
b) Delta Query > SELECY PK_FIELD FROM USERS WHERE LAST_MODIFIED_DATE >
'${dih.last_index_time}'
c) Delta Import Query > SELECT PK_FIELD, JOIN_DATE, USER_NAME FROM USERS WHERE
PK_FIELD = '${dih.delta.PK_FIELD}'
Have a script transformer as below
function dynamicData(){
var joinDt = row.get('JOIN_DATE');
var dtDisplay = joinDt.toString(); //e.g to show that am not doing
null check since join_date is a not null field
...........
...........
return row;
}
<entity name="user" transformer="script:dynamicData" ...... >
.......
</entity>
Problem: While performing delta import, was getting exception from Rhino engine
on the script line 'joinDt.toString()'.
Root Cause: Since I know join_date can not be null, have explored the solr
source code and noticed that applyTransformer() is called during deltaQuery and
at that time join_date will not be available.
Reference: EntityProcessorWrapper.nextModifiedRowKey()
I think transformation is not required for deltaQuery since it is mainly
designed to retrieving the primary keys of the modified rows. Further, the
output of deltaQuery will be used only in another SQL.
Work around:
Just added a null check as a workaround as below
function dynamicData(){
var joinDt = row.get('JOIN_DATE');
if(joinDt == null){
return row;
}
...........
...........
return row;
}
I don't have too much knowledge about Solr and hence my suggestion could be
invalid while looking from main use cases.
Please validate my comments once.
Thanks
Balaji
--
This message is automatically generated by JIRA.
If you think it was sent incorrectly, please contact your JIRA administrators
For more information on JIRA, see: http://www.atlassian.com/software/jira
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
