Your question raises some more questions like....

• How big is file1? If the keys fit into memory, a possible solution could be to read all the John1 => "15.T" type keys into a hash and then parse file2 line by line...
• Does the order of entries in the output file play any role?
• May there be multiple matches and how to treat them?
• Do you count the indexes starting with 0 or 1?
• What is the significance of the dot+character (.T) after the index (15)? E.g., do they select a different file2?
• Is this a one-time task? If not, why not start using a database?
• Why not use a database?
• What have you done so far (show some Perl code or describe your approach)?
• Since this is a common question, did you use Super Search? You might find something like Searching Huge files.
• Why shouting at Perl? It is mostly harmless ;-)
• ...