Hi France,
- (For each //element without @id, add @id) then db:optimise and redirect
to next step 2. For each /*, process content to XLIFF (uses the new ids so requires preceding redirect to commit changes).
Now, is it suggested to work with updindex/autooptimize to true or as shown above... call optimize manually on step1
Both approaches make sense. It clearly depends on your work (the number of affected documents, parallel users, etc.), which one is the best. For example, if it turns out that most of your queries are not rewritten for index access, you may get rid of the explicit optimization. It up-to-date index structures are essential for reasonable performance, and if your data is queried while updates are being performed, UPDINDEX may be the better choice. AUTOOPTIMIZE is just a convenience option, which allows you to get rid of the explicit db:optimize calls.
Hope this helps, Christian