text extractor causing strange character conversion

[Former Member]

Hi experts,

I have created a text extractor for a custom infoobject. The extractor is simple one, reading just one R/3 table. The R/3 table is a text table with this format.

MANDT

SPRAS

custom-field

LGTXT - (data element TEXT40)

One entry in the source table has an entry like this:

"Billing - Costs and Fees"

When the data reaches the BI PSA, that entry becomes this:

"Billing # Costs and Fees"

Note that the '#' (hash sign) is replacing the '-' (dash sign).

I have tested the extractor and the output looks as expected in R/3.

Within BI, there are no conversions on the custom object or in the transfer rules. I have verified that this is not occuring with other text extractors. Has anybody ever seen this?

Signed,

Completely baffled

[Former Member]

After a little 'hex' research on the field in the source system, I found that some invisible disallowed characters were present.