[
https://issues.apache.org/jira/browse/OFBIZ-293?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14907096#comment-14907096
]
Jacques Le Roux commented on OFBIZ-293:
---------------------------------------
I carefully reviewed propositions/comments by Adrian, Nicolas and Paul above
(with corresponding code review) and I must say I agree with Martin. His
proposition is simple, neat and fits in the current implementation of handling
null fields while importing (see r1425813). So I will commit it soon, except if
someone really identifies an issue or propose a better (and still simple ;))
solution.
I though still regret that we have this dichotomy with null and null-field
(nullField) between Entities and Minilang. It makes things confusing and it
would be best if we would have used only one. I guess null, but in my
OFBIZ-1602 comment above I wrote that nullField was introduced because of a
Freemarker issue with null (I did not research again, I guess I was right
then).
At least in Minilang having only one form would definitively clarify this
aspect. I remember having tried to use "null" instead of "nullField" in a
condition-expr and not succeeded when nullField worked. I though can OOTB find
condition-exprS using "null" which now makes me doubt :/. And also why not
using nullField in if-compare. Also we have this difference in screen/form
action? OK, too much for tonight but I feel we are almost there...
> data import - nulling fields impossible
> ---------------------------------------
>
> Key: OFBIZ-293
> URL: https://issues.apache.org/jira/browse/OFBIZ-293
> Project: OFBiz
> Issue Type: Bug
> Components: framework
> Affects Versions: Trunk
> Reporter: Marco Risaliti
> Priority: Minor
> Attachments: OFBIZ-293_set-empty-strings-as-null.patch,
> ofbiz-293-patch.txt, ofbiz-462-patch.txt
>
>
> It's a copy of the old-jira issue
> http://jira.undersunconsulting.com/browse/OFBIZ-462 Tarlika Elisabeth Schmitz
> =================================================
> I am on rev 5274.
> data import from .xml file: empty attributes (e.g. description="") are
> ignored and the field retains its former value.
> Comment by Chris Juettner [04/Oct/05 01:51 PM]
> This patch comments out the check for null or empty values in the data XML
> file found by the EntitySaxReader. I also added an additional warning message
> in case you do not want to persist empty values to the database but still
> would like to know what happened.
> Comment by Si Chen [04/Oct/05 01:59 PM]
> Chris-
> Just a comment: it's very important that a field in the entity engine is set
> to null ONLY when the XML file specifically has an empty attribute (ie,
> description=""). Otherwise, there's a lot of seed data that is in separate
> XML files, and they could overwrite each other.
> Comment by Chris Juettner [10/Oct/05 11:56 AM] [ Permlink ]
> Si,
> I'm not sure I understand your comment about seed data in seperate XML files
> overwriting each other. Shouldn't seed data XML files be entity specific? Why
> would one seed data XML file overwrite another?
> Do you have any comment on what a better solution is for seed data values
> that could be empty?
> Thanks
> Chris
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
