ODS Can't Activate by invalid character

[Former Member]

Hi all,

I have a problem about ODS can't activate.

I have loaded a full request from SAP R/3 into ODS (SAP BW) is completed but my request can't activate. I have tried check error and I met error display as

Value 'ELAMES ' (hex. '454C414D45530D0A2020205B ') of characteristic ZREQUSTER contains invalid character

and then I check error value at PSA that its have ## at this value as "ELAMES ##"

My solution is

1. I deleted request at ODS.

2. go to PSA and deleted ## at error value and then go to RSA1 > PSA > Start update Immediatly to data target

this solution can resolve this problem but its manual solution.

Please help me about better solution for resolve this problem.

Hope its help

Sue.

[Former Member]

Hi all,

If I create ABAP routine at Transfer Rule. How can I do ?

PS. pls describe me about regex ???

Hope you help

Sue.

[Former Member]

Do a search!!!

Here is some regex code I use..

data: w_desc_uc1 type string,
      w_rex type string,
      w_count type i.

w_desc_uc1 = TRAN_STRUCTURE-ITM_DESCRIPT_UC.
w_rex = '[\f\n\r\t\v]'.
* In the above declaration
*\f = form feed, \n = newline, \r = carrige return, \t = horizontal tab,
*\v  = vertical tab

replace all occurrences of regex w_rex in w_desc_uc1  with '|'
REPLACEMENT COUNT w_count.

if w_count > 0.
  do w_count times.
    replace '|' with space into w_desc_uc1.
  enddo.
endif.

  RESULT = w_desc_uc1.

[Former Member]

Hi umesh,

Pls you describe about your solution again. I can't understatnd about its. T.T

hope your help.

Sue.

[Former Member]

Umesh's won;t work as the characters are 0D0A - check the hex and look it up in a translation table

These are carriage return and line feed - ie printer control characters to tell the printer head on a printer what to do!

You need to sort it out at the source system, right a schell scrip using awk or sed to strip them or strip them in the transformation rules using regex or abap code

[Former Member]

Hi all,

I already config # sign in RSKC but when I load data again its still the same error.

Hope you help.

Sue

[Former Member]

Hi,

As mentioned earlier, it not just # value which is causing the error in data load. In BW system all the special chracters and the even the control M characters are displayed as # in the BW backend system. so you need to firstly checek this special characters in your source system and then add those in the RSKC list.

Hope it helps...

Regards,

umesh

[Former Member]

Hi

Do one thing go to your charactics on which u getting error and uncheck alpha check box ,

or run t-code RSKC and put your spetial character and then exe..

points ......

Thanks

Ankit

[Former Member]

Hi,

for invalid characteristics, maintain ALL_CAPITAL in RSKC tcode and execute. Load the data again..

for more info for invalid characteristsics..

hope this helps you ..

Regards,

shikha

[Former Member]

This is quite a simple one

the problem is the 0D0A - those old timers amongst us will recognise this immediately as a carriage return / line feed

ie you have printer control characters stuck in your string

The SAP dynpro as mentioned in the second post cannot identiofy which ascii character to display for this so diplays hash

There are numerous ways of changing this in code if you cannot trap this at your front end systems

If this is a falt file you could ask yoru basis guys to maybe put a awk or sed script together to strip them

Otherwise you can use regex in your transformation rules if you are using the uptodate basis patch

Else you can loop around using one of the many code samples mentioned above

(and no - you cannot add this to RSKC these are control characters - you need to strip them)

[Former Member]

Hi,

Check out the below thread:

Vivek

[Former Member]

Hi,

This is common production support issue while doing ETL for BW, as all the special chracters are detected as # in BW system. If the issue is occasional then you can go ahead with the solution you've mentioned above.

Other method to overcome this issue is by firstly identifying this characters at the sources system level in extraction. alternate approach is to write a Program which will run through the PSA records and erase such special characters from PSA, so the request is properly loaded.

Hope it helps..

Regards,

Umesh

[Former Member]

Hi,

fo this you load that characteristic through PSA .. and there you edit the erroneous records ..and then again try activating the .

the ODS ..this is the temprory solution to procced with loads ..

but you have to fix the invalid characteristics ..in source system..

hope this helps you ...

reagrds,

Shikha

Edited by: Shikha Goel on Oct 7, 2008 7:03 AM