I've had the same challenge and solved it with a tab separated rule file. It could look something like this:
TRANCODE QUANTITY QSIGN KEYCODE HANDLER
/(TSP|KSP)/ /^.*(?<!0000000.00000)$/ + /(10|11|20|60)/ se
+t(TRAN_CODE, "RCV")
/(TSP|KSP)/ * - /(10|11|20|60)/ se
+t(TRAN_CODE, "DLV")
...
The table is meant as a lookup table where each value (apart from the values in the rightmost field) are lookup keys that are processed left to right like this:
- /<some regex>/ matches against <some regex>
- * matches against anything
- All other values are considered literals and must match exactly (see the plus and minus signs above)
The different field types have different precedence, so a * will have lower precedence than a regex, which in turn has lower precedence than a literal value.
The rightmost field is just a value that says what to do when matched. It can be a reference to some handler or just a value to insert somewhere.
I do have a module to do this kind of lookup, but it is currently not open source. If you really want it, I can ask the owners if they will let me release it.
|