COBOL has had XML parsing for a while, and has generation. Of late it uses the 
z/OS System Service. For JSON parsing (parsing only so far) it also uses the 
existing z/OS System Service.

So Enterprise COBOL has native parsing/generation for XML and native parsing 
for JSON. It has nothing native for CSV.

Probably for anything new JSON would be better, followed by XML, followed by 
CSV, followed by fixed-width fields (are they really so difficult for others to 
process?).

It is true that fields in CSVs have no "type", but an Excel Macro, which runs 
from a button, to load a specific CSV, and which checks the date (being for the 
expected business date, not "run-date") and logical file-name on the "header", 
and totals one or more numeric fields, and counts the data-records, and 
compares them to a trailer, and backs up the previous spreadsheet, and renames 
and saves the source files (so that a previous situation can be recovered, or 
new functionality in the spreadsheet can be run with prior data), can mean that 
the spreadsheet already knows what the formats are (and the data is known to be 
for the correct day, the correct file, and is unlikely to have 
lost/gained/changed records and is not an incomplete file) gets around that 
issue and also has the other stated benefits. These are what I remember from 
'91.

----------------------------------------------------------------------
For IBM-MAIN subscribe / signoff / archive access instructions,
send email to [email protected] with the message: INFO IBM-MAIN

Reply via email to