Your translation of RudiC's suggestion was close... I think you would have gotten what you had expected if the 2nd line marked in red in your code above had been:

Code:

                        {p1=$7

instead of:

Code:

                        {p1=$9

The code Aia suggested should work fine as long your data never contains two people with the same name with different IDs. RudiC's code was matching on the ID (e.g., 1111, 2222, and 3333) while Aia's code is matching on the name (e.g., John Smith, Elane Boozler, and Hortense Gable). With the sample data you provided you could have matched on any field except $3, $9, and $10; but I assume that ID ($7) and name ($8) are the only two reasonable choices to consider for your real data.

Note also that the awk statement delete array_name (i.e., delete TMP) is an extension to the standards and will not work on some awk implementations. If you ever need to port this to a system that doesn't accept that statement, the following should work on any standards conforming version of awk:

Code:

                         for(i in TMP) delete TMP[i]