L3 - Default Context:
Currently we have index: true for xml fields of type: text, we need it indexed as we perform exists queries to obtain definitions without an xml yet.
However this comes at the cost that text fields are by default tokenized and analyzed, we don't need and want this overhead. We should disable at least tokenization on xml fields but still manage to store the information whether the xml field is empty or not.
AT:
- xml fields are only analyzed whether they are empty or not (as we use exists query) no tokenization is perfomed (which is done by default on text fields)