Importing/exporting data with carriage returns (I Think)

[Brian K. White]

On 8/28/2017 3:16 PM, Fairlight via Filepro-list wrote:
> They call it -logic-. ...

I don't see that Mike did anything wrong deserving of these responses.

He's not a programmer and lacked the background to make sense of your 
initial answer. That is not illegal. Once you establish that, just 
switch into that mode.


Mike, this kind of thing can be addressed several possible ways, and 
Mark's first answer was really just meant as a general description or 
basic outline of one possible approach, not any kind of exact commands 
or details. No one could offer any exact details yet at that point 
because we didn't yet know enough details about your situation. First 
comes picking a possible approach, then comes hashing out what exact 
details are required to make that happen.

That idea might not have been practical for you for some reason, or 
might be doable but might not be the most convenient option available. 
It was too early at that point to do more than essentially spitball some 
different basic possible approaches.

For instance:

* If there are CR's or LF's or both within the cells, then how ARE the 
records delimited, if not by CR? Maybe you don't need to do anything but 
adjust the record delimiter option on the import command line in 
processing. That would be much better than cobbling together some 
external pre-processing steps.

* Or, you could do your own more manual parsing of the data by using 
open and readline instead of import. Then you could detect incomplete 
records and read in the next line, and concatenate to the previously 
read line, repeat until you know you have read in a complete record (by 
counting the commas that weren't inside any quotes or something). This 
would not be convenient to write at all, but it's one and only advantage 
is, it would be done entirely in filepro processing and doesn't depend 
on any external programs or batch files, and once it's done, it's done. 
It'll then 'just work" forever.

* Or, if you control the export, then maybe you could alter the export 
to make it possible for fp to import it reliably. (again, even this one 
idea, has many possible meanings. Impossible to guess them before even 
knowing the first thing: "if you control the export". Only IF that is 
true, the next question would be, what program are you exporting from? 
maybe that program has some option to specify the record delimiter, or 
if not, maybe you can add another column to the data that you could then 
look for in processing to detect the real end of the record. Or maybe 
you could export in a fixed-length format, and then the filepro import 
wouldn't even look at any LF or CR at all. They would just be bytes like 
any other bytes.)

* Or maybe pre-processing the data is the most practical way to go after 
all.

-- 
bkw