Offloading NVivo files

Part 3

So the first thing I wanted to do was fix up those strangely coded fields. After a fair bit of messing around I worked out that all that had happened was that the hex value 0x377 had been added to the Unicode value of each of the characters. So I wrote a quick Python script Translate to undo that mess. To use it, just type:

python Translate <name of your SQLite file>

Note also that must only do this once - otherwise the decoding will be re-applied and you're fields will be messed up in the opposite direction.

So now you have a reasonably tidy SQLite mirror of your NVivo file that you can manipulate using your tools and operating system of preference.

Below are some notes I made looking over the tables and fields, which may be of some use.

Item.ItemID

0 - Folders
2 - Source (internal)
3 - Source (Memo)
4 - Source (external)
34 - Source (Online)
14 - ? Query/Report something?
16 - Node
17 - Node Matrix
24 - Matrix Coding Query
21 - Attribute value
51 - Source classifications
52 - Node classifications
20 - Attribute name

Role.TypeId

0 - -
1 - Item2 is Node ID

2 - Attribute parent-child - Item1 is parent, Item2 is child, tag=1 means with aggregation
6 - Attribute name - Item1 is attribute name, Item2 Attribute value, Tag is value ordinal
7 - Source attribute value -  Item1 is source, Item2 Attribute value
13 - Attribute - Item1 is attribute name, Item2 is Source classification, Tag is ordinal?
14 - Source/Node Category - Item1 is Source/Node Item2 is Category
15 - Item2 is Node, Item1 is Parent Node for inherited coding

Written on May 27, 2015