We have source as sequential file. Followed by seqential file stage we have transformer stage and then we have ODBC stage (say A) followed by one more ODBC stage (say B).
Between stage A and B we have reject link. Records which gets rejected due to primary key constraint or duplicacy gets automatically inserted into reject table. The problem with us is that the structure of original table and reject table differs by one field. The original table and the reject table, both have 10 columns. Name of column 2 in original table is "status_code" and in reject table it is "status_value". To correct this we have kept a modifier stage in between stage A and B. Due to this reject link (dotted line) has gone and the column definitions of reject stage B went off. Is this possible to modify the column name by keeping modifer stage in between stage A and B? Will it change the reject link properties.
problem with reject link
Moderators: chulett, rschirm, roy
If the so called 'reject table' was designed specifically for this job then I would say standardize the meta data at the table level as well. In otherwords, change the column name to whats in the file. This way there is conformity of metadata.
Creativity is allowing yourself to make mistakes. Art is knowing which ones to keep.