Data manipulation

Yogiz · 12-11-2007, 07:21 PM

Hi there,

I have a bit of a problem with data manipulation below is the output, would it be possible to use a joint command or anything else to produce output file where the numbers in the middle columns mach each other like the first record in the input file.

Code:

Input
=============================
0.918	100	100	1.32
0.918	101	102	1.328
0.926	102	104	1.336
0.926	103	106	1.344
0.934	104	108	1.352
0.934	105	110	1.36
0.942	106	112	1.368
0.942	107	114	1.376
0.95	108	116	1.384
0.95	109	118	1.392
0.958	110	120	1.4
0.958	111	122	1.408
0.966	112	124	1.416
0.966	113	126	1.424
0.974	114	128	1.432
0.974	115	130	1.44
0.982	116	132	1.448
0.982	117	134	1.456
0.99	118	136	1.464
0.99	119	138	1.472
0.998	120	140	1.48

Output
==============================
0.918	100	100	1.32
0.918	101
0.926	102     102	1.328
0.926	103
0.934	104
0.934	105
0.942	106     106	1.344

Thanks in advance.

pixellany · 12-11-2007, 08:41 PM

I can't see the logic based on what you have shown. At first I thought you meant "write the line to output if cols 2 and 3 are the same".

You will need to describe the logic (algorithm) in more detail.

sundialsvcs · 12-11-2007, 09:00 PM

There's more than one way to do it, obviously. (Hope I don't sound too much like a Perl-conversant programmer when I say that.) And there are several very powerful tools you can consider.

For example, in any Unix/Linux environment you can use the notion of "piping the output of one program into another" to do things like sorting your input and then processing it. Lots of programs know how to "accept input from STDIN and send it to STDOUT."

One very powerful tool is awk (see: man tt) which is specifically designed for dealing with text files. It is based on a disarmingly-simple premise that is astonishingly useful and powerful.

A step up from that is Perl, a genuine programming-language that is somewhat of "awk on super-steroids."

Start by considering your problem: describe it in terms of an algorithm that can be applied in order to solve it. In Linux/Unix, it is not a question of lacking for the tools to get the job done! The question is merely, how you wish to go about doing it.