Laboratoire ICAR » Plateforme TXM

Currently, empty or missing word properties are encoded by default by '__UNDEF__' in cwb-encode.

The user should be able to choose the code.

Solution¶

Create a new global Preference (in the Import section) named 'Empty property value code' (emptyPropertyValueCode for insiders).

Validation test¶

• import a short corpus with XML/w or XTZ import module
  

  <test>
  <w t="1">very</w>
  <w t="">small</w>
  <w t="2"></w>
  <w t="3">small</w>
  </test>
  

• set the empty property value code to "EMPTY"
• the "word" lexicon must show:
  

      small    2
      EMPTY    1
      very    1
  

• the "t" lexicon must show:
  

      1    1
      2    1
      3    1
      EMPTY    1