DSXchange

The option for Preserve partioning - default(clear) has been set.

Where has it been set? It needs to be set to 'Clear' on the stage before the sequential stage ie on the transformer.

I make that an average of > 7 posts a day (including holidays and weekends, naturally).

When do you sleep?

Well done, here's to the next 10K ...

Phil Clarke.

Thanks Arnd - So the short answer is 'No'. I should have mentioned that Null Field Value was set to ''. This is our default. By inconsistent I meant that an empty input column is treated differently depending on the nullability of the target column. The (online) documentation defines the Null Field ...

Given a pipe-delimited input sequential file something like this Row 1 Data1|Data2|Data3 Row 2 Data1||Data3 All fields, Varchar(6), say. What would you expect to happen to row 2 if Field 2 was defined as (a) Nullable and (b) Not nullable. The answer seems to be that (a) Datastage inserts a Null. (b)...

Thanks for your reply....we can check for the duplicte rowids before updating and if there are no duplicates then we can go ahead. Is that OK? You don't need to check for duplicates, ROWIDs will always be unique within a table. You only need to be certain the ROWID will not change between reading a...

When I said 'query' the db design I was using the word in the sense of to question or challenge the design of the table. It is one of the fundamental principles of relational table design that there should be an attribute in the table that uniquely identifies a row. http://en.wikipedia.org/wiki/Uniq...

This will probably work in this scenario, but using ROWID may fail in some circumstances. Oracle ROWID is a psuedo-column, determined by the RDBMS rather than the data. It is unique within a table and is fixed in most but not all circumstances. For example exporting and reimporting a table will chan...

Another way to skin the Vertical Pivot cat... Sort and add a KeyChange column as in previous suggestions. Add a transformer that increments a stage variable and resets on key change giving a Counter field in the output: If KeyChange=1 then svCOunt = 1 Else svCount = svCount + 1 Add a switch stage wi...

My understanding, based on a document from IBM, is that where I have a Left Outer join and a row in the 'Left' dataset does not find a match in the 'Right' dataset, any nullable columns from the right dataset will be nulled. But this does not seem to be the case for timestamps. I have a logical test...

In that case the setting of APT_STRING_PADCHAR may well be the problem. This specifies the character that Datastage will use to pad out strings when you move a string to a longer CHAR field - for example a 4-character string to a CHAR(6) - it will add two of this character to the end of the field. T...

Agree. Remember also that routine names are case-sensitive; I had the same error and traced it back to routine names spelled the same but with different capitalisation.

Ooops

So you did. In case of conflicting advice, readers should believe Arnd over me every time

Here is the storage required for each data type. Using these you can get a pretty good estimate of your dataset record size... Data Type: Size Integers: 4 bytes Small Integer: 2 bytes Tiny Integer: 1 byte Big Integer: 8 bytes Decimal: (precision+1)/2, rounded up Float: 8 bytes VarChar(n): n + 4 byte...

Thanks Ray. That simple, huh? Is this documented anywhere?

DSGuru - the question you answered was not the question I asked.

What is the difference between 'Set' and 'Propagate' in the 'Preserve Partitioning' Drop-Down box (usually on the Advanced Tab)? A colleague encountered the 'Partitioning despite the Preserve Paritioning Flag' warning. I advised to change the Propagate option to 'Clear', he pointed out that he wante...