Sequential file Import problem

[somu_june]

Hi,

HAPPY NEW YEAR

I am facing an import problem when I am trying to read a tab delimited file . When I am running the job, sequential file stage is dropping the records with the following warning


Sequential_File_7,0: Field "SDATE" delimiter not seen, at offset: 27.

I have input data like this in the file

"T#PSR4003 "||Jan 1, 0001||"40"

Where first one is CSE char 11 and second one is SDATE char 12

and in sequential file stage properties I defined properties like this

Record level

Final delimiter = none
Record delimiter = UNIX newline

Field defaults
Delimiter String = ||
quote = double

please help me in solving my problem

Thanks
Soma Raju

[ray.wurlod]

Only single-character delimiters are supported. You will need to change your file, or process it as a single field and decompose it using Field() functions in a Transformer stage. There will be empty fields between adjacent pipe characters that you can ignore.

[narasimha]

Other options...

If you don't have any pipes in your data (other than the double pipe delimiter),
You could preprocess the file, replace all the '||' with '|'

Another option could be to read the file as is with '|' as the delimiter.
Then in your job you can map only the required fields and ignore the rest.

I am facing an import problem when I am trying to read a tab delimited file .

Clarification
Is your file a tab delimited or a '||' delimited?

[just4geeks]

Other options...

If you don't have any pipes in your data (other than the double pipe delimiter),
You could preprocess the file, replace all the '||' with '|'

By preprocessing, do you mean using awk? Or something in DataStage? Just wanted to check if I got what you meant.

[narasimha]

[By preprocessing, do you mean using awk? Or something in DataStage? Just wanted to check if I got what you meant.

Not sure how you are related to the OP.
But yes you are right, thats what I meant.
You could do it outside Datastage or build it inside Datastage, your choice.
You can also use sed, which every you are comfortable in.

[naren]

to my understanding CSE char 11 is "T#PSR4003 "
and SDATE char 12 Jan 1, 0001

if thats the case try varchar instead of char, and add one extra column in metadata at the end and use varchar 3 and let me know whether u got the answer.

[DSguru2B]

I think sed should be enough.
Something like

Code: Select all

sed 's/||/|/' myFile.txt

This searches all occurances of '||' and replaces it with '|'.

[ray.wurlod]

Missing a "g"?

[kumar_s]

What Ray and DSguru2B is suggesting is, use

Code: Select all

sed 's/||/|/g' File.txt > Newfile.txt

either in BeforJobSubroutine or using a Execute command in JobSequence where the job been called, and read the Newfile.txt for you lay out.

[DSguru2B]

Missing a "g"?

Was missing it , Kumar filled in the space....Thanks Kumar.

[somu_june]

Hi,

My problem was solved and in input data the length of the date field is not constant , for some it has 14 char length and for others it has 12 char length. After making it as constant length, I can load the data with out dropping the records





Thanks,
SomaRaju

[narasimha]

Thanks for sharing your solution with everybody.
Now I guess its time to mark your post as resolved
If possible can give more details on what path you choose, what delimiter you used,....