It is important to realize that the extraction process does involve a true SQL join between the rule columns of the parent and child tables in the relationship. Like any join, if the certain columns are not indexed the execution can be slow. In the case of a DataBee extract it is quite important that the columns in the child table in the relationship are indexed otherwise there could be performance problems. It is not important in terms of extraction performance whether or not the columns in the parent table are indexed.
The Unindexed Rules Detector dialog box searches through the rules in the extraction set and identifies the rules in which the target columns in the child table are not indexed. This allows you to make decisions about whether to discard the rule, add a temporary index or perhaps to find an alternative rule to populate the child table.
In the illustration above one rule has been identified as having unindexed columns in the child table. In this case the DTB_INVOICE_LINE table has no index on its INVOICE_NUMBER column. It is not important if the DTB_INVOICE table has an index on its corresponding INVOICE_NUMBER column.
Note: It is not always necessary to drop or do anything about a rule which uses unindexed columns in the child table. It simply means that that particular rule will extract slower - it will always extract correctly. The slower extraction speed may not even be noticeable if the table is small or may not be a problem if you are prepared to wait for the rule to complete.