uima-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Jasper Huzen (JIRA)" <...@uima.apache.org>
Subject [jira] [Created] (UIMA-5777) Incorrect feature assignment on MARKTABLE because incorrect record can be used
Date Wed, 16 May 2018 12:59:00 GMT
Jasper Huzen created UIMA-5777:
----------------------------------

             Summary: Incorrect feature assignment on MARKTABLE because incorrect record can
be used
                 Key: UIMA-5777
                 URL: https://issues.apache.org/jira/browse/UIMA-5777
             Project: UIMA
          Issue Type: Bug
          Components: Ruta
    Affects Versions: 2.6.1ruta
            Reporter: Jasper Huzen


Feature assignment with MARKTABLE can go wrong. 

Let assume that we have a CSV table with the following entries:
||Matching||ID||
|First Item|1|
|Second Item|2|
|SECOND ITEM|3|

and we use the MARKTABLE action so that we match on the first column and assign the second
column to a feature of the annotaion.

If we do the match case sensitive and we use the following input as document: "SECOND ITEM"
the annotation will get "2" as feature value.

The MARKTABLE action match correct on "SECOND ITEM". In the next step it tries to set the
features. Therefore it call the getRowWhere method on the (csv) table to get the related
row in the table.

The code in the getRowWhere method always compare the tableValue and lookup value lowercase
and that will result in the record with "Second Item" in it. That's incorrect because we need
the one with  "SECOND ITEM". 

I modified the code so that it first start with an exact match. If no matching item is found
and we ignore case it will also do a case insensitive check.



--
This message was sent by Atlassian JIRA
(v7.6.3#76005)

Mime
View raw message