WOCAT » QCAT

h1. QT QA Migration

h2. 1. Process Overview

There are different steps take, same for technology as for approaches:

#  *Mapping*  between the new questionnaire in qcat.wocat.net and the old questionnaire in qt.wocat.net

** This mapping says which attribute/question in qcat takes the information from which attribute in qt

**  It explains the transformation, which has to be done from the old attribute to the new one.

#  *Migration script* which 

** extracts the data from the old database (qt.wocat.net)

** transforms the data based on the rules defined in the mapping table

** imports the case in the new database (qcat.wocat.net) in draft mode

#  *Post-editing* all imported cases

** Adding missing data by coping from qt what was not automatically imported.

** Reviewing the cases and publishing them to make them available publically.

h2. 2.  Mapping between old and new questionnaire

This mapping is needed for the two following actions:

* For the development of the migration script. 

* For the post-editing in qcat to find the missing data in the old qt. 

Therefore it needs the mapping in two forms:

Once with the qcat questions as index/keys and once with the qt questions as index/keys...

h2. Migration script / Extract-Transform-Insert

h3. How it works

1. Migration script migrates all cases and their attributes automatically, as long as¨..

* transform doesn't produce errors because of inconsistent or wrong data

** Example: If a value should have been selected only once but was then selected twice. -> the script produces an error

* transformation can be done automatically 

** Example: In the old questionnaire, several values could be selected, in the new one only one. -> Transform can't be made automatically.

2. In cases of errors or the impossiblity of automatic transform:

* Not migrated or problematic questions (in qcat) are listed per case id in a migration log.

3. All cases are inserted as 'submitted for review' in qcat. 

h2. 3. Post-editing of imported cases in qcat

* The post-editing person needs two things: 

** The migration log 

** The mapping table

* Post-editing tasks:

** Adding the missing data from qt.wocat.net follwoing the migration log and using the mapping table

** Publishing the cases

*Question:* When should the compiler be informed about the transfer? 

*Suggestion:*  When the case is published on qcat.