I have a flat file containing person "Records" which I get from a service provider. Each Record is in fact a list of Records. Each line in the file represents a Record with a given Type. A person Record always contains the types '001' and '099', representing a start- and an end Record fora person. Other records are added to a person entry based on an 'incident'. Hence a Record is built based on the given incident.
Example.
000 # Start of doc.
001 # Person begin.
...
n Records (n > 001 && n < 099)
...
099 # Person end.
001 # Person begin.
...
n Records (n > 001 && n < 099)
...
099 # Person end.
001 # Person begin.
...
n Records (n > 001 && n < 099)
...
099 # Person end.
999 # End of doc.
Apart from the Record Type on each line there are other data contained within fixed positions.
Example of actual start Record
0010308952872 01000000000000 K1995-03-01
My approach, so far, is to define an .XSD per type of record and use schema binding classes. My problem is that there are still no requirements in regards to what types of records are relevant based on the type of incident. I would just grab a phone and call the person responsible, but working for the public sector makes planning a b****.
Does anyone have any suggestions to which software pattern could adhere most to this type of problem statement? Any sharing of experience would be much appreciated :}
Aucun commentaire:
Enregistrer un commentaire