Sponsored Content
Top Forums Shell Programming and Scripting Sort based on Multiple Columns in UNIX Post 302833265 by Chubler_XL on Tuesday 16th of July 2013 04:15:02 PM
Old 07-16-2013
Sorry, should have checked my sort parameters the 4th field was supposed to be sorted in descending order to do the sort needs the r flag (not d) so your correct answers are:
Code:
sort -t\; -k1,1 -k4,4r infile | awk -F\; '!($1 in A) { print $0; A[$1] }'
sort -t\; -k1,1 -k2,2 -k4,4d infile | awk -F\; '!($1$2 in A) { print $0; A[$1$2] }'

For the awk logic:
-F\; use semicolon as the field separator, this allows awk to split the line up into $1, $2 ... $NF fields

!($1$2 in A) means only process lines where the value of field1 concatenated with field2 (your unique key) is not already in the array A

print $0 print the entire line

So in prints the first line where each unique key value occurs.

A[$1$2] add value field1 concatenated with field2 to the array A
This User Gave Thanks to Chubler_XL For This Post:
 

10 More Discussions You Might Find Interesting

1. UNIX for Dummies Questions & Answers

Help needed to sort multiple columns in one file

Hi, I would like to know given that I have 3 columns. Let say I have first 3 columns to do operation and these operation output is printed out each line by line using AWK and associative array.Currently in the output file, I do a sort by -r for the operation output. The problem comes to... (1 Reply)
Discussion started by: ahjiefreak
1 Replies

2. Shell Programming and Scripting

Remove lines, Sorted with Time based columns using AWK & SORT

Hi having a file as follows MediaErr.log 84 Server1 Policy1 Schedule1 master1 05/08/2008 02:12:16 84 Server1 Policy1 Schedule1 master1 05/08/2008 02:22:47 84 Server1 Policy1 Schedule1 master1 05/08/2008 03:41:26 84 Server1 Policy1 ... (1 Reply)
Discussion started by: karthikn7974
1 Replies

3. Shell Programming and Scripting

sort by multiple columns and reformat...

hello.. I have big file and so far I was able to shink it and make smaller with certains values that I need.. vendor1|2000|1 vendor2|1000|1 vendor2|5000|1 vendor2|500|2 vendor3|1000|2 vendor3|500|3 vendor4|500|3 Vendor5|500|1 vendor6|500|3 Vendor7|1000|1 Vendor8|1000|774... (3 Replies)
Discussion started by: abdulaziz
3 Replies

4. Shell Programming and Scripting

sort by based on multiple columns

Hi, Is there any way to sort a file in cshell by sort command, sorting it by multiple fields, like to sort it first by the second column and then by the first column. Thanks forhead (1 Reply)
Discussion started by: Takeeshe
1 Replies

5. Shell Programming and Scripting

sort on multiple columns

Howdy! Need to sort a large .txt file containing the following, using sort. First based on the 1st column, and then on the 2nd column: Group01.01 1000500 31 0.913 -1.522974494 Group01.01 1001500 16 0.684 -0.967496041 Group01.01 36500 19 0.476 na Group01.02 365500 15 0.400 na... (1 Reply)
Discussion started by: sramirez
1 Replies

6. Shell Programming and Scripting

sort on multiple columns

hi all, i have a file , having few columns. i wanted to sort it based on 2nd column and then based on 1st column. But i have some problem in first column. first column have characters and numbers, but problem is number of characters are not same in all rows. Few rows have 13 characters and then... (3 Replies)
Discussion started by: deepakiniimt
3 Replies

7. Shell Programming and Scripting

Sort and join multiple columns using awk

Is it possible to join all the values after sorting them based on 1st column key and replace empty rows with 0 like below ? input a1 0 a1 1 a1 1 a3 1 b2 1 a2 1 a4 1 a2 1 a4 1 c4 1 a3 1 d1 1 a3 1 b1 1 d1 1 a4 1 c4 1 b2 1 b1 1 b2 1 c4 1 d1 1 output... (8 Replies)
Discussion started by: quincyjones
8 Replies

8. UNIX for Dummies Questions & Answers

UNIX Sort on a field that spans multiple columns

New to unix. I need to sort the records of a file by a control number field. That field is in POS 16 through 28. How do I do that? There are no delimiters, or spaces to separate fields. See example below. Each line is a record. REC1CCYYMMDD0018888888888888ABCDE... (1 Reply)
Discussion started by: jclanc8
1 Replies

9. Programming

Ls -ltr Sort multiple columns

Hi All, I have one requirement, where I need to have output of ls -l command sorted on 1) first on filename 2) last modified time ( descending ) - latest change first. I am not able to figure out how to do it.. Also I dont have a way to change Date display for ls -ltr command.. I am... (1 Reply)
Discussion started by: freakabhi
1 Replies

10. UNIX for Beginners Questions & Answers

Sort and remove duplicates in directory based on first 5 columns:

I have /tmp dir with filename as: 010020001_S-FOR-Sort-SYEXC_20160229_2212101.marker 010020001_S-FOR-Sort-SYEXC_20160229_2212102.marker 010020001-S-XOR-Sort-SYEXC_20160229_2212104.marker 010020001-S-XOR-Sort-SYEXC_20160229_2212105.marker 010020001_S-ZOR-Sort-SYEXC_20160229_2212106.marker... (4 Replies)
Discussion started by: gnnsprapa
4 Replies
ICON(1) 						      General Commands Manual							   ICON(1)

NAME
icon - interpret or compile Icon programs SYNOPSIS
icont [ option ... ] file ... [ -x arg ... ] iconc [ option ... ] file ... [ -x arg ... ] DESCRIPTION
icont and iconc each convert an Icon source program into executable form. icont translates quickly and provides interpretive execution. iconc takes longer to compile but produces programs that execute faster. icont and iconc for the most part can be used interchangeably. This manual page describes both icont and iconc. Where there there are differences in usage between icont and iconc, these are noted. File Names: Files whose names end in .icn are assumed to be Icon source files. The .icn suffix may be omitted; if it is not present, it is supplied. The character - can be used to indicate an Icon source file given in standard input. Several source files can be given on the same command line; if so, they are combined to produce a single program. The name of the executable file is the base name of the first input file, formed by deleting the suffix, if present. stdin is used for source programs given in standard input. Processing: As noted in the synopsis above, icont and iconc accept options followed by file names, optionally followed by -x and arguments. If -x is given, the program is executed automatically and any following arguments are passed to it. icont: The processing performed by icont consists of two phases: translation and linking. During translation, each Icon source file is translated into an intermediate language called ucode. Two ucode files are produced for each source file, with base names from the source file and suffixes .u1 and .u2. During linking, the one or more pairs of ucode files are combined to produce a single icode file. The ucode files are deleted after the icode file is created. Processing by icont can be terminated after translation by the -c option. In this case, the ucode files are not deleted. The names of .u1 files from previous translations can be given on the icont command line. These files and the corresponding .u2 files are included in the linking phase after the translation of any source files. The suffix .u can be used in place of .u1; in this case the 1 is supplied auto- matically. Ucode files that are explicitly named are not deleted. iconc: The processing performed by iconc consists of two phases: code generation and compilation and linking. The code generation phase produces C code, consisting of a .c and a .h file, with the base name of the first source file. These files are then compiled and linked to produce an executable binary file. The C files normally are deleted after compilation and linking. Processing by iconc can be terminated after code generation by the -c option. In this case, the C files are not deleted. OPTIONS
The following options are recognized by icont and iconc: -c Stop after producing intermediate files and do not delete them. -e file Redirect standard error output to file. -f s Enable full string invocation. -o name Name the output file name. -s Suppress informative messages. Normally, both informative messages and error messages are sent to standard error output. -t Arrange for &trace to have an initial value of -1 when the program is executed and for iconc enable debugging features. -u Issue warning messages for undeclared identifiers in the program. -v i Set verbosity level of informative messages to i -E Direct the results of preprocessing to standard output and inhibit further processing. The following additional options are recognized by iconc: -f string Enable features as indicated by the letters in string: a all, equivalent to delns d enable debugging features: display(), name(), variable(), error trace back, and the effect of -f n (see below) e enable error conversion l enable large-integer arithmetic n produce code that keeps track of line numbers and file names in the source code s enable full string invocation -n string Disable specific optimizations. These are indicated by the letters in string: a all, equivalent to cest c control flow optimizations other than switch statement optimizations e expand operations in-line when reasonable (keywords are always put in-line) s optimize switch statements associated with operation invocations t type inference -p arg Pass arg on to the C compiler used by iconc -r path Use the run-time system at path, which must end with a slash. -C prg Have iconc use the C compiler given by prg ENVIRONMENT VARIABLES
When an Icon program is executed, several environment variables are examined to determine certain execution parameters. Values in paren- theses are the default values. BLKSIZE (500000) The initial size of the allocated block region, in bytes. COEXPSIZE (2000) The size, in words, of each co-expression block. DBLIST The location of data bases for iconc to search before the standard one. The value of DBLIST should be a blank-separated string of the form p1 p2 ... pn where the pi name directories. ICONCORE If set, a core dump is produced for error termination. ICONX The location of iconx, the executor for icode files, is built into an icode file when it is produced. This location can be overridden by setting the environment variable ICONX. If ICONX is set, its value is used in place of the location built into the icode file. IPATH The location of ucode files specified in link declarations for icont. IPATH is a blank-separated list of directories. The current directory is always searched first, regardless of the value of IPATH. LPATH The location of source files specified in preprocessor $include directives and in link declarations for iconc. LPATH is otherwise sim- ilar to IPATH. MSTKSIZE (10000) The size, in words, of the main interpreter stack for icont. NOERRBUF By default, &errout is buffered. If this variable is set, &errout is not buffered. QLSIZE (5000) The size, in bytes, of the region used for pointers to strings during garbage collection. STRSIZE (500000) The initial size of the string space, in bytes. TRACE The initial value of &trace. If this variable has a value, it overrides the translation-time -t option. FILES
icont Icon translator iconc Icon compiler iconx Icon executor SEE ALSO
The Icon Programming Language, Ralph E. Griswold and Madge T. Griswold, Prentice-Hall Inc., Englewood Cliffs, New Jersey, Second Edition, 1990. Version 9.1 of Icon, Ralph E. Griswold, Clinton L. Jeffery, and Gregg M. Townsend, IPD267, Department of Computer Science, The University of Arizona, 1995. Version 9 of the Icon Compiler, Ralph E. Griswold, IPD237, Department of Computer Science, The University of Arizona, 1995. icon_vt(1) LIMITATIONS AND BUGS
The icode files for the interpreter do not stand alone; the Icon run-time system (iconx) must be present. Stack overflow is checked using a heuristic that is not always effective. 1 November 1995 IPD244b ICON(1)
All times are GMT -4. The time now is 11:22 AM.
Unix & Linux Forums Content Copyright 1993-2022. All Rights Reserved.
Privacy Policy