File split question

I have a flat file in UNIX and I have to perform two tasks based on the below data. The data I have printed here is just sample the original data is too long.

The position 110 to 111 (two digit value I have bolded the values) theygives the record type detail in the sample above the record types in the sample are 32,32,31,31 and 35. The real data contains thousands of more records and there are more than 100 record types in a file. I have to split the file based on the record types in position 110 THRU 112.

000000008101 000011000700000000000000000000000001234567454002000 832I20090109 1234567097009967
123450007101 000000000000000000007856343446560000007856454540000 832I20090109 9864536670002456
957645465778 000011000700000000000000000000000067645333567743355 831I20090109 7854536670005647
676767497101 000011000700000000000000000000000008898675335767676 831I20090109 4565767665545469
767865444567 000011000700000000000000000000000007876564454676877 835I20090109 8786756656677887

TASK1: I have to split the file based on the record types. So the out put in this case will be three files

File1 RecordType32
000000008101 000011000700000000000000000000000001234567454002000 832I20090109 1234567097009967
123450007101 000000000000000000007856343446560000007856454540000 832I20090109 9864536670002456

File2 RecordType31
957645465778 000011000700000000000000000000000067645333567743355 831I20090109 7854536670005647
676767497101 000011000700000000000000000000000008898675335767676 831I20090109 4565767665545469

File3 RecordType35
767865444567 000011000700000000000000000000000007876564454676877 835I20090109 8786756656677887

Can any body help me with a solution for this? I am not good at UNIX shell scripting

TASK2: I need to get a unique list of record types in a file in my sample the result should be
32
31
35

Man this is pretty easy. Surprised no one followed up:

awk 'length($0) > 111 { type=substr($0,110,2); ofile="type-" type ".dat"; print $0 > file; } 
       length($0) <= 111 { print $0 >"type-short.dat" }'

Hi otheus,
Thank you for your reply. I appreciate it. I will try this at work tomorrow as I do not have access fro home.

Have a nice day

Assuming the record types are in the 67th position as in your example and not in position 110, this should be sufficient:

awk '{print > "RecordType" substr($0,67,2)}' file

Regards

Is the part I highlighted in red a typo? Was it supposed to be ofile instead of type? I didn't test it- just seemed so...

For task 2- I believe you can use the sed command or awk command, piped to the sort command with the -u option. Google search for "sed" and for "unix sort" should probably turn up lots of good info for you to learn.

zTodd is correct. the Red 'type' should have been 'ofile' (since corrected). Thanks...

Task 1 and 2 in one go:

awk '{_types[substr($3,2,2)]; print > "/tmp/type"substr($3,2,2)} END {for (i in _types) print i }' file  > /tmp/type.list

The /tmp/type.list file is not sorted. If you need to sort:

awk '{_types[substr($3,2,2)]; print > "/tmp/type"substr($3,2,2)} END {for (i in _types) print i}' file | sort >  /tmp/type.list

If your version of awk supports the asorti() function you could also use it but I don't recommend its use as it is rather cumbersome. (In fact it is the function I hate the most in gawk!)