DSXchange

A signed zero is a very unusual thing. EBCDIC '0C' is packed decimal +0, a rather common thing in many mainframe data repositories. Data attributes that include an explicit sign often require +0 as the default value. This is true in Cobol standards (when they are followed). Of course, unsigned zero...

There are some details you haven't given us -- not complaining, just mentioning -- that might help decide on the right answer. I have had no problems letting DataStage handle character sets. I read EBCDIC and write ASCII without any explicit conversions in between. 1) The data type doesn't tell me w...

In addition to what Craig wrote, using a third-party scheduler means you can create dependencies between DataStage and other runtime engines, or, as in my case, multi-platform schedules using Windows and mainframe.

You may be a bit confused here: integer is a storage format, but you are describing a requirement for displaying the data. The easiest way to handle this is to convert the integer to a char or varchar before sending it to your display output. The only alternative is to store it in the display format...

Please don't ask me for specifics -- I am not in a position to share code, regretfully -- but this is something to which I have had to give some thought. Your first step is to flowchart your Cobol programs. I have found that, in the absence of a utility for conversion, a flowcharted program can poin...

In general, trying to read EBCDIC mainframe files is actually simpler than one might expect. Try the following settings for your read format: Record level>>Record type=implicit Field defaults>>Delimter=none Type defaults>>General>>Character set=EBCDIC, Data format=binary Type defaults>>Decimal>>allo...

Yeah, IBM owns Lotus. I'll suppress further commentary on that. :twisted: I see two problems: Can you be sure the attachment will always be in the correct OS format? Seems to me that Lotus runs under the Windows/DOS environment (unless your shop runs it on Unix?), and files will generally be stored ...

We have a similar situation, and the control point for setting the parameters for each instance is a table. It can be set dynamically, in that the number of jobs executed is equal to the number of rows in the parameters table.

My direct DW experience is with 7x and Oracle (source data is DB2 or sequential datasets), so the following is a general comment. 1) To what use will the DW tables be put? If, for example, it's reporting of some kind, consider designing the fact tables with partitions that support the reporting peri...

It took a few days after I paid for Premium. Pay no attention to the man behind the curtain, the Great Oz has spoken!

I deal primarily with COBOL datasets that have multiple record types in addition to header and trailer, so I've given some thought to what would be either easiest or least confusing. Ideally, your header and trailer records would have as many delimiters as your detail records. That is usually unlike...

The one setting you don't list is Record level>>Record Type=implicit. I've not tried that with a delimiter, but my read of mainframe data works just fine on any table definition I've used so far so long as it conforms to the LRECL setting. Each row fits nicely into the columns and implied widths the...

After further thought, it occured to me that this issue should start in the requirements. Any amount that is intended to be used as a denominator should have error handling for zero before it gets to the calculation. In COBOL, this is a standard item for an entire paragraph that does nothing but val...

I am calculating a column like this Column A = (Column B+Column C)/Column D Now if suppose a zero value comes at Column D then the Datastage will through an exception ..or an error so how can I handle this.. ???? Standard approach: Insert If condition before doing calculation that defaults Column A...

Ha, Ray! My COBOL instructor way back when was the IT directory for a small book publisher. There's a "liberal" and an "art" in there for the asking.