DSXchange

Ken once made the comment that Red Brick users are spoiled with their bulk loader. He's right. Another way that they're spoiled is that the bulk loader can be made to update the aggregate tables automatically! [8D] (Off the same data set that is used to load the main tables.) Ray Wurlod Education an...

Parameters are properties of job objects (which are instances of the job class). So the only methods that are available are DSGetParamInfo and DSSetParam, or their implicit or inherited equivalents in Job Activities in job sequences or in stage or pin objects. Because of the way that DataStage is de...

The array of fields is one-based, not zero-based; that is, the first field number is 1, not zero. DataStage BASIC has a shorthand notation called "delimited substring extraction" that you may prefer; though it generates exactly the same compiled code as the FIELD function. For example: Field001 = Re...

Truncating the hashed file will probably gain you nothing, since all writes to hashed files are destructive overwrites. The usual way to "truncate" the hashed file is to check the "clear file before writing" check box on the hashed file stage Inputs tab. However, if you'd like to try it as a command...

For DataDirect drivers visit their website: m (These are the drivers currently shipped with DataStage.) For Merant drivers visit their website: m (Are these guys still in the driver business?) For OpenLink drivers visit their website: m (An alternative.) Ray Wurlod Education and Consulting Services ...

As Kim says, it's a matter of which is attempted first. Note, though, that these are both inefficient in that there will be failed attempts made against the database in either case. Best practice is to split the stream of data into those rows that need to be inserted (don't already exist in the data...

CREATE.INDEX does not generate any output. Note, too, that this command only creates the structure; if there are already data in the hashed file, you will need to build the index. The LIST.INDEX command will report whether there are any secondary indices defined for the file. CREATE.INDEX filename c...

"Not supported" means exactly that.

Try declaring the column in DataStage as VarChar, and (if necessary) using used-defined SQL to CAST the column AS VARCHAR.

SELECT ..., CAST (COLUMN_TEXT AS VARCHAR(255)) FROM ...

Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518

Cleanup project among other things rebuilds the indexes in the repository database. Since everything in the repository is indexed on CATEGORY, this explains why cleaning up the project fixes the problem; the indexing had either not been built properly or had become out of synchronization. Ray Wurlod...

Number of developers is licensed separately from the number of telnet users. In version 6 most things that used to be uv... are now ds..., so you need to use dslictool.exe. Syntax is: dslictool report_lic (There is another option, clean_lic -a, to clean up after users become defunct.) At version 6.0...

You have two choices. One is to code using BCI functions, which mimic ODBC functions. To do this from a UNIX-based DataStage server, you will need to have licensed your ODBC driver. The other is to construct a small job that executes the SQL statement via ODBC and deposits the result in, for example...

vmdsapi.dll is the Windows object that exposes the functions documented for the DataStage Job API. If you have NLS enabled, you will see names that include the language, for example vmdsapiENU.dll for English (US) or vmdsapiJPN.dll for Japanese. uvclnt32.dll does not exist for DataStage 6.0.1; it is...

What you see in Director as the status of the job is a column from the RT_STATUSnn table for that job. This is updated periodically by the running job. If you kill the job, the "status" column can not be updated by the running job. This is why the job appears still to be running. This is also why ev...

I think you'll be surprised how fast the UNION (or UNION ALL) approach works, provided you don't include an ORDER BY clause in it.

The second one. All writes to hashed files are destructive overwrites of the entire row.


Ray Wurlod
Education and Consulting Services
ABN 57 092 448 518