DSXchange

lna_dsuser wrote:Source contains only 1 row

I'm assuming that means always. The required output won't make much sense with multiple rows.

A non-DataStage solution? Google for "convert dbf to csv" to find alternatives.

Akilasaminathan1 wrote:DBASE is not a source, the .DBF files from DBASE is source.

Same thing. The file is the database. Use the ODBC Connector stage after creating an ODBC data source name using the dBase drivers.

Alt+Ctl, Alt+Shift and Ctl+Shift are all easily reached with the left hand and then you punch G with the right. Trying to keep all three on the left hand would mean using Q or S to get something... comfortable... to press but nothing over in that area makes sense like G does. Same goes for the right...

Out of curiousity, are you allowed more than a 'two key' combination? Wondering about something like Ctl-Alt-Del... err, I mean Ctl-Alt-G. Seems like that may get it more 'out of the way' of other applications.

I meant it to mean "un-hijacked"... as in split out to its own post and linked back to the original post it came from. Got nothing to do with the problem or solution, just a Moderator thing related to the fact that one shouldn't really jump on an old post... if you have a problem and can't...

ray.wurlod wrote:(once I'd understood your # to belong to "Row" and not to "=").

'Not' being the operative word here.

I'm thinking you would handle the column names the same way... dynamically paired with the matching non-zero values. It would need to be written out as a row of data rather than automatically by the stage, so perhaps a horizontal pivot? Or a funnel to bring the two rows together.

The syntax is in the documentation. Does your version support an 'optional format string'? That would simplify things but from some searches here and online 8.1 might not allow that optional third parameter.

That ampersand needs to be replaced with & to be valid in XML data.

Or read it as a single field, strip out the double quotes and then let a Column Export stage create the two fields.

That's close to what I was thinking. :wink: It doesn't need to be read as one column so much as written out that way. And I don't think the Column Import stage is needed. As the final step I would trim off the 'extra' trailing delimiter you'd get from concatenating the columns and delimiters togethe...

Then I would suggest you involve your official support provider, perhaps there is a patch available.

What version are you on? Regardless - did you check the documentation for IsValid? It is multi-purpose as you might gather from the name: Returns whether the given string is valid for the given type. Valid types are "date", "decimal", "dfloat", "sfloat", "...