Re: CRTCSVF vs CPYTOIMPF

rob@xxxxxxxxx wrote:

Well, if you are distributing software then CRTCSVF is not on my systems and that could be an issue.

Yeah, CRTSAVF is a third-party package.

I was so disgusted with IBM on V5R3 and up what they did with CPYTOIMPF and CPYFRMIMPF that I won't use them anymore. I find that writing my own is:
- faster
- more granular on handling errors.
- doesn't blow me grief on CCSID crud.

I know how much you hate CPYxIMPF, so here's an interesting option, Rob:

MySQL supports a CSV type. There are a number of issues that limit the usefulness of that storage method (no primary keys, for example), but it seems pretty well suited to import/export processing. Of course you're limited to whatever MySQL thinks is appropriate for formatting, but it seemed pretty reasonable - double quotes for alpha fields, that sort of thing.

So step one, if you're doing something on a regular basis, you could write a simple Java program to read from DB2 and write to the CSV file. Yeah, I know this is a little Rube Goldberg, but here's the thing: MySQL AB is supposedly working with IBM to provide a native DB2 engine for MySQL. I don't know what the status of this project is, but if it happens, you could theoretically install MySQL on the i and create MySQL definitions for both your CSV file and your DB2 file. You could then use the MySQL GUI to run ad hoc queries to extract data from DB2 and insert it into the CSV file (and vice versa).

Yeah, I know, it's a lot of futures and finger crossing, but as a potential strategic direction it may have some merit.

If you want to see a couple of pictures including the query browser (I even show dumping data from a traditional MySQL table into a CSV file), I wrote about installing and running MySQL here:

http://www.mcpressonline.com/database/mysql/mysql-and-ibm-i-where-does-mysql-fit-in-a-db2-world.html

Joe