High/low value identify

techsavy · Post by **techsavy** » Tue May 05, 2015 1:16 pm

Hi,

I have a ebcdic file which I am reading through CFF Stage. There is a column where high and Low values are expected. I need to reject those records. Can I identify high/low value in datastage?

Need your suggestions

FranklinE · Post by **FranklinE** » Tue May 05, 2015 2:27 pm

First, that sounds very unusual. "Filling" on high-value (xFF) is useful only when the field is used in a sort and they want all such records at the end of the sort. Low_value is x00, which is also the ASCII null. Having either or both in a field doesn't sound right at all, and if it were me I'd ask for the justification for it.

Settings to reject such records should be relatively easy except for the low-value/null problem. You can examine each byte and set to reject upon finding x00 or xFF. Position-length reference on a Char or VarChar field is setting at the end of the column name [pos,length].

Good luck.

techsavy · Post by **techsavy** » Tue May 05, 2015 10:42 pm

Thanks Franklin,

Can you give an example to identify low(x00) and high(xFF) value in datastage. I am just wondering how to catch this hex representation in datstage.

ray.wurlod · Post by **ray.wurlod** » Tue May 05, 2015 11:42 pm

Char(0) and Char(255) respectively.

But this only matches a single character.

What do you really want to accomplish?

techsavy · Post by **techsavy** » Wed May 06, 2015 6:45 am

Hi Ray,

I want to identify those records in transformer stage where the specific column having high or low value.

FranklinE · Post by **FranklinE** » Wed May 06, 2015 6:54 am

I believe Ray joins me in asking for further details, something specific about the column being examined.

Will it be all x00 or all xFF? Will it be some mix of the two?

The coding for this is basic. You examine the entire column or some portion of it. And to echo the usual question here: what have you tried so far, and what were the results?

chulett · Post by **chulett** » Wed May 06, 2015 7:16 am

techsavy,

Please make use of the lovely Reply to topic button, there's no need to be quoting everything you are replying to all the time. The DSXchange Janitorial Service thanks you.

techsavy · Post by **techsavy** » Wed May 06, 2015 7:52 am

The field will either only be high value or only low value (not both).

Incase high value - All the position of the field will be filled with xFF simillary for low value - All the position filled with x00

chulett · Post by **chulett** » Wed May 06, 2015 7:56 am

So perhaps simply substring off the first or last character and check that.

FranklinE · Post by **FranklinE** » Wed May 06, 2015 2:24 pm

techsavy wrote:The field will either only be high value or only low value (not both).

Incase high value - All the position of the field will be filled with xFF simillary for low value - All the position filled with x00

Example: transformer constraint

Code: Select all

inlink.colname[1,1] <> Char(00) And inlink.colname[1,1] <> Char(255)

Edit: negative logic has to use "and". Also, you can use a Filter stage and if needed create a reject link for the "bad" records.

techsavy · Post by **techsavy** » Wed May 06, 2015 9:02 pm

Thanks Franklin. I think I can write similar
logic to Check all positions instead of
Only first position. Right?

ray.wurlod · Post by **ray.wurlod** » Wed May 06, 2015 10:05 pm

Probably easier is to use Convert() function to eliminate the low values and high values, then see if there's anything left.

Code: Select all

Convert(Char(0):Char(255),"",InLink.TheField) > ""

chulett · Post by **chulett** » Wed May 06, 2015 10:11 pm

techsavy wrote:Thanks Franklin. I think I can write similar logic to Check all positions instead of only first position. Right?

Why if they're always all of one or the other? I suggested checking one character because that's what you stated. But...sure, if you feel the need.

techsavy · Post by **techsavy** » Thu May 07, 2015 5:13 am

I tried with Convert but not sure why its not working. When I see the data in datastage view mode it's look like '-1' in place of high value.

chulett · Post by **chulett** » Thu May 07, 2015 6:59 am

Post your derivation code and let us know the data type of the field you are using it on.