# how to model string pattern to character occurence in deep learning

In my question on stackoverflow, I was trying to decode concatenated string into partitions for example,

SQCB7A750BATWE SQ CB 7 A 750 B A T WE
PT05A1219PY023 PT 05 A 12 19 P Y 023
PT55A1019PX02  PT 55 A 10 19 P X 02
PT33SE2215SW023    PT 33 SE 22 15 S W 023
PT05A2216PW023(LC) PT 05 A 22 16 P W 023 (LC)


my first thought was to use seq2seq model but, it didn't work out. Am trying to think about the problem differently now, am thinking of modeling a pattern instead of values to whitespace" " occurences for example

PT05A1219PY023 PT 05 A 12 19 P Y 023 pair


is converted to

11001000011000  [2,5,7,10,13,15,17] pair


where alphabets have been converted to 1 and numericals have been converted to 0 (might have alphanumerics too) and [2,5,7,10,13,15,17] represents "whitespace" occurences.

my question is would such model makes sense ? or there is a better way to solve the problem.