WinMergeX Line Transformations

Line Transformations are a way to normalize input data files on the fly.

To consider an example, we are comparing many data dumps, and some of them use US-style dates (eg, 3/14/2005), and others use (ISO) standard dates (eg, 2005-03-14).

We want to see these lines compare as equal:

Left	Right
3/14/2005: $1400	2005-03-14: $1400

but these lines compare as different:

Left	Right
3/14/2005: $1400	2005-03-14: $1500

To solve this problem using WinMergeX, we must use Line Transformations.

We need a transform map, which maps input files to named line transformations; for this case we will map files inside any subdirectory named "data" to the line transform "Date Standardizer".
- DataDate.TransformMap
We need a definition of the line transform "Date Standardizer".
- DateStandardizer.LineTransform
Place both of these files (the transform map and the line transform) inside either of your WinMergeX filters directories, and WinMergeX will find and parse them upon startup.

The main portion of the actual mechanical process that occurs is the line transformation engine testing a regular expression such as this "(\d)/(\d\d)/(\d\d\d\d)" against a line, and when it matches, doing a substitution with replacement of captured substrings, like this: "{$3}-0{$1}-{$2}".

Here we captured the three parts of a US-date, reordered them, and added a leading zero to the one-digit day number. (In the actual sample line transform file above, four substitutions were used, to handle both one and two digit months and both one and two digit day numbers.)

Home