Extract is picking up rows twice because both an explicit TABLE and a TABLE

Question

Our extract parameter file has this following(full file uploaded):
TABLE TK_APP_DATA.PAYMENT_INFO , COLSEXCEPT (ENCRYPTED_VALUE,HASH_VALUE)
#add_scn()
TABLE TK_APP_DATA.*
#add_scn()
When i look at the change data inserted by the replicat on target and at the extract trail file I can see that a row which is updated is actually extracted twice and duplicated. I also see that the table is resolving twice in the EXTRACT rpt file:
Using the following key columns for source table TK_APP_DATA.PAYMENT_INFO : INSTRUMENT_ID, FIELD_HANDLE.
TABLEWildcard resolved (entry SCECOMM.*):
TABLE SCECOMM.PAYMENT_INSTRUMENT_FIELD, TOKENS ( TK-SCN = @GETENV ( "TRANSACTION" , "CSN" ) ) ;
Using the following key columns for source table TK_APP_DATA.PAYMENT_INFO : INSTRUMENT_ID, FIELD_HANDLE.
TABLEWildcard resolved (entry SCECOMM.*):
TABLE SCECOMM.APPLICATION_PARAMETER, TOKENS ( TK-SCN = @GETENV ( "TRANSACTION" , "CSN" ) ) ;
Anyone know how I can make it so that only one TABLE statement is used to resolve it? Or conversely a way to filter out table in a wild card statement? I would prefer to avoid having to create a seperate extract for just 1 table(schema has over 350+ tables in it)
I tried using a TABLEEXCLUDE but that excludes the table entirely. I also tried using WILDCARDRESOLVE IMMEDIATE on the specific table mappings, but can't because we also use DDL replication.
Any help would be greatly appreciated.
#anh

Answer 1

This is unfortunately what I have right now as my stop gap. We will sql generate the list of tables with every deployment which with our agile development is like every other week haha. I do have a requirement to replicate dynamically(the source side changes extremely frequently by adding and dropping tables) but even after talking to oracle support this is not possible. If you have a custom mapping for a table and use a wild card later which would resolve that table you will always get duplicate rows extracted.
I talked to oracle support(with raised this as a bug/feature request) and raised this with our oracle engagement manager and asked them to add better functionality when it comes to filtering and wild card selection for TABLE and MAP statements to have an option setting to only map once. It's in their queue however not sure how high of a priority it will be for them(I'm guessing not very high =) )