DSXchange

Load todays file in TEMP1 and yesterdays file in TEMP2. Once you finish your change detection, truncate TEMP2 and load it with TEMP1's data. Finally truncate TEMP1. Tomorrow when you get fresh file, load it into TEMP1. This way your TEMP1 will always have todays file and TEMP2 will always have yeste...

Well, if its a column by column comparison then kcbland already gave one suggestion; to load the files to temp tables and do join. Another way would be to load one file to a hashed file keyed on all the columns and do a lookup against the other file. reflink.NOTFOUND will give you the changes in the...

Yes, True, I forgot that these numeric values can change. Thanks for jumping in Craig.

Look at Craig's 3'rd reply in thispost.

Specify the output as varchar and see if the following works for you: Specify a stage variable say stgVar. Let its derivation be DecimalToString(in.Field). For the derivation of the particular column, specify the following: If INDEX(stgVar, "-", 1) > 0 then stgVar else "+"...

Never saw that error. Whats your design like? I think its a better idea to get in touch with support as not very many of us have had a chance to work with 8.0

ICONV/OCONV wont work in parallel jobs.
First you need to seperate records which have date length less than 10. The ones that pass, apply StringToDate() function to them.

You have to recognize a field as a key in order to detect any change.

If you dont mind using the transformer you could do the following in the constraint

This will give you all records whose date length is less that 10.

How many stages do you have? If its less, copy paste them to a different job, instead of copying the job or doing 'Select All'.

Can't these guys edit a file and put the parameter values in the file?

Go to the home directory

or

and type uv or dssh. This will take you to the uv prompt.

Go into datastage help, go to the last tab Search and type Parallel Job Functions , hit 'List Topics'. It will give you a bunch of topics, click on Parallel Job Functions . It will give you a list on the righ hand side of all the parallel functions. Click on them and find them in the chart. These ar...

What do you mean its truncating? Is it truncating the data (can you confirm by doing a head at the unix level) or truncating the length? If you want to honor the length then specify the field as char.

Thanks for coming back with the resolution.