[thelist] importing a non-standard datadump

Luther, Ron ron.luther at hp.com
Wed Feb 23 10:49:47 CST 2005


Brian Cummiskey noted:

>>oh, no no.  I still have the orig. data file.  I'm not *that* dumb :p


Hi Brian!


Coolio!! 

(I hope you know I didn't mean to offend -- that's a lesson everybody 
has to learn the hard way once. [Okay, more than once for some of us!] 
Good to hear that one didn't trip you up!)  ;-)


>>I think the number of rows i have is correct.  But one, or more of 
>>these rows is missing a column.  I've tryed to skim it a couple 
>>times to see if it would stand out, and its just impossible.

Okay.  How 'bout breaking the original file into smaller chunks?

-- Run 1,000 lines of data through your processing instead of all 
24,000 and see what happens. That okay? Then try 3,000. You can 
interpolate to find one of the troublemakers.

-- Mix your own 'dummy file' containing 5 'good', 5 'bad', and then 
5 more 'good' records. Run it through your processing and see if all 
15 records turn out as intended.

Some playing around like that should help you isolate one of the 
situations the processing is stuttering on. 

Then you can fix it!


>>Uglier than this- I hope I never have to see that.

Heh. Imagine a couple thousand page mainframe 'green bar' report 
"printed" to disk. Page numbers. Field headers on every page. 
'Blank' lines for 'visual' spacing. Variable number of records. 
Variable number of sections. Different fields in each section. 
Different sections needing to be parsed to different tables. Some 
fields from 'section 1' needing to be 'remembered' so you can write 
them out with 'section 3' in order to link the tables back 
together for reporting later ... It can be quite a 'fun' project!


Good Luck!

RonL.


More information about the thelist mailing list