Mercurial > repos > devteam > merge_cols
changeset 0:28ca7552e884
Uploaded merge_cols tarball.
| author | devteam |
|---|---|
| date | Tue, 04 Dec 2012 11:00:46 -0500 |
| parents | |
| children | f2aac0c5c60d |
| files | mergeCols.py mergeCols.xml test-data/1.bed test-data/mergeCols.dat |
| diffstat | 4 files changed, 230 insertions(+), 0 deletions(-) [+] |
line wrap: on
line diff
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/mergeCols.py Tue Dec 04 11:00:46 2012 -0500 @@ -0,0 +1,37 @@ +import sys, re + +def stop_err( msg ): + sys.stderr.write( msg ) + sys.exit() + +def __main__(): + try: + infile = open ( sys.argv[1], 'r') + outfile = open ( sys.argv[2], 'w') + except: + stop_err( 'Cannot open or create a file\n' ) + + if len( sys.argv ) < 4: + stop_err( 'No columns to merge' ) + else: + cols = sys.argv[3:] + + skipped_lines = 0 + + for line in infile: + line = line.rstrip( '\r\n' ) + if line and not line.startswith( '#' ): + fields = line.split( '\t' ) + line += '\t' + for col in cols: + try: + line += fields[ int( col ) -1 ] + except: + skipped_lines += 1 + + print >>outfile, line + + if skipped_lines > 0: + print 'Skipped %d invalid lines' % skipped_lines + +if __name__ == "__main__" : __main__() \ No newline at end of file
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/mergeCols.xml Tue Dec 04 11:00:46 2012 -0500 @@ -0,0 +1,63 @@ +<tool id="mergeCols1" name="Merge Columns" version="1.0.1"> + <description>together</description> + <command interpreter="python"> + mergeCols.py + $input1 + $out_file1 + $col1 + $col2 + #for $col in $columns + ${col.datacol} + #end for + </command> + <inputs> + <param format="tabular" name="input1" type="data" label="Select data" help="Dataset missing? See TIP below."/> + <param name="col1" label="Merge column" type="data_column" data_ref="input1" /> + <param name="col2" label="with column" type="data_column" data_ref="input1" help="Need to add more columns? Use controls below."/> + <repeat name="columns" title="Columns"> + <param name="datacol" label="Add column" type="data_column" data_ref="input1" /> + </repeat> + </inputs> + <outputs> + <data format="tabular" name="out_file1" /> + </outputs> + <tests> + <test> + <param name="input1" value="1.bed"/> + <param name="col1" value="4" /> + <param name="col2" value="1" /> + <param name="datacol" value="6" /> + <output name="out_file1" file="mergeCols.dat"/> + </test> + </tests> +<help> + +.. class:: infomark + +**TIP:** If your data is not TAB delimited, use *Text Manipulation->Convert* + +----- + +**What it does** + +This tool merges columns together. Any number of valid columns can be merged in any order. + +----- + +**Example** + +Input dataset (five columns: c1, c2, c3, c4, and c5):: + + 1 10 1000 gene1 chr + 2 100 1500 gene2 chr + +merging columns "**c5,c1**" will return:: + + 1 10 1000 gene1 chr chr1 + 2 100 1500 gene2 chr chr2 + +.. class:: warningmark + +Note that all original columns are preserved and the result of merge is added as the rightmost column. + </help> +</tool>
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/test-data/1.bed Tue Dec 04 11:00:46 2012 -0500 @@ -0,0 +1,65 @@ +chr1 147962192 147962580 CCDS989.1_cds_0_0_chr1_147962193_r 0 - +chr1 147984545 147984630 CCDS990.1_cds_0_0_chr1_147984546_f 0 + +chr1 148078400 148078582 CCDS993.1_cds_0_0_chr1_148078401_r 0 - +chr1 148185136 148185276 CCDS996.1_cds_0_0_chr1_148185137_f 0 + +chr10 55251623 55253124 CCDS7248.1_cds_0_0_chr10_55251624_r 0 - +chr11 116124407 116124501 CCDS8374.1_cds_0_0_chr11_116124408_r 0 - +chr11 116206508 116206563 CCDS8377.1_cds_0_0_chr11_116206509_f 0 + +chr11 116211733 116212337 CCDS8378.1_cds_0_0_chr11_116211734_r 0 - +chr11 1812377 1812407 CCDS7726.1_cds_0_0_chr11_1812378_f 0 + +chr12 38440094 38440321 CCDS8736.1_cds_0_0_chr12_38440095_r 0 - +chr13 112381694 112381953 CCDS9526.1_cds_0_0_chr13_112381695_f 0 + +chr14 98710240 98712285 CCDS9949.1_cds_0_0_chr14_98710241_r 0 - +chr15 41486872 41487060 CCDS10096.1_cds_0_0_chr15_41486873_r 0 - +chr15 41673708 41673857 CCDS10097.1_cds_0_0_chr15_41673709_f 0 + +chr15 41679161 41679250 CCDS10098.1_cds_0_0_chr15_41679162_r 0 - +chr15 41826029 41826196 CCDS10101.1_cds_0_0_chr15_41826030_f 0 + +chr16 142908 143003 CCDS10397.1_cds_0_0_chr16_142909_f 0 + +chr16 179963 180135 CCDS10401.1_cds_0_0_chr16_179964_r 0 - +chr16 244413 244681 CCDS10402.1_cds_0_0_chr16_244414_f 0 + +chr16 259268 259383 CCDS10403.1_cds_0_0_chr16_259269_r 0 - +chr18 23786114 23786321 CCDS11891.1_cds_0_0_chr18_23786115_r 0 - +chr18 59406881 59407046 CCDS11985.1_cds_0_0_chr18_59406882_f 0 + +chr18 59455932 59456337 CCDS11986.1_cds_0_0_chr18_59455933_r 0 - +chr18 59600586 59600754 CCDS11988.1_cds_0_0_chr18_59600587_f 0 + +chr19 59068595 59069564 CCDS12866.1_cds_0_0_chr19_59068596_f 0 + +chr19 59236026 59236146 CCDS12872.1_cds_0_0_chr19_59236027_r 0 - +chr19 59297998 59298008 CCDS12877.1_cds_0_0_chr19_59297999_f 0 + +chr19 59302168 59302288 CCDS12878.1_cds_0_0_chr19_59302169_r 0 - +chr2 118288583 118288668 CCDS2120.1_cds_0_0_chr2_118288584_f 0 + +chr2 118394148 118394202 CCDS2121.1_cds_0_0_chr2_118394149_r 0 - +chr2 220190202 220190242 CCDS2441.1_cds_0_0_chr2_220190203_f 0 + +chr2 220229609 220230869 CCDS2443.1_cds_0_0_chr2_220229610_r 0 - +chr20 33330413 33330423 CCDS13249.1_cds_0_0_chr20_33330414_r 0 - +chr20 33513606 33513792 CCDS13255.1_cds_0_0_chr20_33513607_f 0 + +chr20 33579500 33579527 CCDS13256.1_cds_0_0_chr20_33579501_r 0 - +chr20 33593260 33593348 CCDS13257.1_cds_0_0_chr20_33593261_f 0 + +chr21 32707032 32707192 CCDS13614.1_cds_0_0_chr21_32707033_f 0 + +chr21 32869641 32870022 CCDS13615.1_cds_0_0_chr21_32869642_r 0 - +chr21 33321040 33322012 CCDS13620.1_cds_0_0_chr21_33321041_f 0 + +chr21 33744994 33745040 CCDS13625.1_cds_0_0_chr21_33744995_r 0 - +chr22 30120223 30120265 CCDS13897.1_cds_0_0_chr22_30120224_f 0 + +chr22 30160419 30160661 CCDS13898.1_cds_0_0_chr22_30160420_r 0 - +chr22 30665273 30665360 CCDS13901.1_cds_0_0_chr22_30665274_f 0 + +chr22 30939054 30939266 CCDS13903.1_cds_0_0_chr22_30939055_r 0 - +chr5 131424298 131424460 CCDS4149.1_cds_0_0_chr5_131424299_f 0 + +chr5 131556601 131556672 CCDS4151.1_cds_0_0_chr5_131556602_r 0 - +chr5 131621326 131621419 CCDS4152.1_cds_0_0_chr5_131621327_f 0 + +chr5 131847541 131847666 CCDS4155.1_cds_0_0_chr5_131847542_r 0 - +chr6 108299600 108299744 CCDS5061.1_cds_0_0_chr6_108299601_r 0 - +chr6 108594662 108594687 CCDS5063.1_cds_0_0_chr6_108594663_f 0 + +chr6 108640045 108640151 CCDS5064.1_cds_0_0_chr6_108640046_r 0 - +chr6 108722976 108723115 CCDS5067.1_cds_0_0_chr6_108722977_f 0 + +chr7 113660517 113660685 CCDS5760.1_cds_0_0_chr7_113660518_f 0 + +chr7 116512159 116512389 CCDS5771.1_cds_0_0_chr7_116512160_r 0 - +chr7 116714099 116714152 CCDS5773.1_cds_0_0_chr7_116714100_f 0 + +chr7 116945541 116945787 CCDS5774.1_cds_0_0_chr7_116945542_r 0 - +chr8 118881131 118881317 CCDS6324.1_cds_0_0_chr8_118881132_r 0 - +chr9 128764156 128764189 CCDS6914.1_cds_0_0_chr9_128764157_f 0 + +chr9 128787519 128789136 CCDS6915.1_cds_0_0_chr9_128787520_r 0 - +chr9 128882427 128882523 CCDS6917.1_cds_0_0_chr9_128882428_f 0 + +chr9 128937229 128937445 CCDS6919.1_cds_0_0_chr9_128937230_r 0 - +chrX 122745047 122745924 CCDS14606.1_cds_0_0_chrX_122745048_f 0 + +chrX 152648964 152649196 CCDS14733.1_cds_0_0_chrX_152648965_r 0 - +chrX 152691446 152691471 CCDS14735.1_cds_0_0_chrX_152691447_f 0 + +chrX 152694029 152694263 CCDS14736.1_cds_0_0_chrX_152694030_r 0 -
--- /dev/null Thu Jan 01 00:00:00 1970 +0000 +++ b/test-data/mergeCols.dat Tue Dec 04 11:00:46 2012 -0500 @@ -0,0 +1,65 @@ +chr1 147962192 147962580 CCDS989.1_cds_0_0_chr1_147962193_r 0 - CCDS989.1_cds_0_0_chr1_147962193_rchr1- +chr1 147984545 147984630 CCDS990.1_cds_0_0_chr1_147984546_f 0 + CCDS990.1_cds_0_0_chr1_147984546_fchr1+ +chr1 148078400 148078582 CCDS993.1_cds_0_0_chr1_148078401_r 0 - CCDS993.1_cds_0_0_chr1_148078401_rchr1- +chr1 148185136 148185276 CCDS996.1_cds_0_0_chr1_148185137_f 0 + CCDS996.1_cds_0_0_chr1_148185137_fchr1+ +chr10 55251623 55253124 CCDS7248.1_cds_0_0_chr10_55251624_r 0 - CCDS7248.1_cds_0_0_chr10_55251624_rchr10- +chr11 116124407 116124501 CCDS8374.1_cds_0_0_chr11_116124408_r 0 - CCDS8374.1_cds_0_0_chr11_116124408_rchr11- +chr11 116206508 116206563 CCDS8377.1_cds_0_0_chr11_116206509_f 0 + CCDS8377.1_cds_0_0_chr11_116206509_fchr11+ +chr11 116211733 116212337 CCDS8378.1_cds_0_0_chr11_116211734_r 0 - CCDS8378.1_cds_0_0_chr11_116211734_rchr11- +chr11 1812377 1812407 CCDS7726.1_cds_0_0_chr11_1812378_f 0 + CCDS7726.1_cds_0_0_chr11_1812378_fchr11+ +chr12 38440094 38440321 CCDS8736.1_cds_0_0_chr12_38440095_r 0 - CCDS8736.1_cds_0_0_chr12_38440095_rchr12- +chr13 112381694 112381953 CCDS9526.1_cds_0_0_chr13_112381695_f 0 + CCDS9526.1_cds_0_0_chr13_112381695_fchr13+ +chr14 98710240 98712285 CCDS9949.1_cds_0_0_chr14_98710241_r 0 - CCDS9949.1_cds_0_0_chr14_98710241_rchr14- +chr15 41486872 41487060 CCDS10096.1_cds_0_0_chr15_41486873_r 0 - CCDS10096.1_cds_0_0_chr15_41486873_rchr15- +chr15 41673708 41673857 CCDS10097.1_cds_0_0_chr15_41673709_f 0 + CCDS10097.1_cds_0_0_chr15_41673709_fchr15+ +chr15 41679161 41679250 CCDS10098.1_cds_0_0_chr15_41679162_r 0 - CCDS10098.1_cds_0_0_chr15_41679162_rchr15- +chr15 41826029 41826196 CCDS10101.1_cds_0_0_chr15_41826030_f 0 + CCDS10101.1_cds_0_0_chr15_41826030_fchr15+ +chr16 142908 143003 CCDS10397.1_cds_0_0_chr16_142909_f 0 + CCDS10397.1_cds_0_0_chr16_142909_fchr16+ +chr16 179963 180135 CCDS10401.1_cds_0_0_chr16_179964_r 0 - CCDS10401.1_cds_0_0_chr16_179964_rchr16- +chr16 244413 244681 CCDS10402.1_cds_0_0_chr16_244414_f 0 + CCDS10402.1_cds_0_0_chr16_244414_fchr16+ +chr16 259268 259383 CCDS10403.1_cds_0_0_chr16_259269_r 0 - CCDS10403.1_cds_0_0_chr16_259269_rchr16- +chr18 23786114 23786321 CCDS11891.1_cds_0_0_chr18_23786115_r 0 - CCDS11891.1_cds_0_0_chr18_23786115_rchr18- +chr18 59406881 59407046 CCDS11985.1_cds_0_0_chr18_59406882_f 0 + CCDS11985.1_cds_0_0_chr18_59406882_fchr18+ +chr18 59455932 59456337 CCDS11986.1_cds_0_0_chr18_59455933_r 0 - CCDS11986.1_cds_0_0_chr18_59455933_rchr18- +chr18 59600586 59600754 CCDS11988.1_cds_0_0_chr18_59600587_f 0 + CCDS11988.1_cds_0_0_chr18_59600587_fchr18+ +chr19 59068595 59069564 CCDS12866.1_cds_0_0_chr19_59068596_f 0 + CCDS12866.1_cds_0_0_chr19_59068596_fchr19+ +chr19 59236026 59236146 CCDS12872.1_cds_0_0_chr19_59236027_r 0 - CCDS12872.1_cds_0_0_chr19_59236027_rchr19- +chr19 59297998 59298008 CCDS12877.1_cds_0_0_chr19_59297999_f 0 + CCDS12877.1_cds_0_0_chr19_59297999_fchr19+ +chr19 59302168 59302288 CCDS12878.1_cds_0_0_chr19_59302169_r 0 - CCDS12878.1_cds_0_0_chr19_59302169_rchr19- +chr2 118288583 118288668 CCDS2120.1_cds_0_0_chr2_118288584_f 0 + CCDS2120.1_cds_0_0_chr2_118288584_fchr2+ +chr2 118394148 118394202 CCDS2121.1_cds_0_0_chr2_118394149_r 0 - CCDS2121.1_cds_0_0_chr2_118394149_rchr2- +chr2 220190202 220190242 CCDS2441.1_cds_0_0_chr2_220190203_f 0 + CCDS2441.1_cds_0_0_chr2_220190203_fchr2+ +chr2 220229609 220230869 CCDS2443.1_cds_0_0_chr2_220229610_r 0 - CCDS2443.1_cds_0_0_chr2_220229610_rchr2- +chr20 33330413 33330423 CCDS13249.1_cds_0_0_chr20_33330414_r 0 - CCDS13249.1_cds_0_0_chr20_33330414_rchr20- +chr20 33513606 33513792 CCDS13255.1_cds_0_0_chr20_33513607_f 0 + CCDS13255.1_cds_0_0_chr20_33513607_fchr20+ +chr20 33579500 33579527 CCDS13256.1_cds_0_0_chr20_33579501_r 0 - CCDS13256.1_cds_0_0_chr20_33579501_rchr20- +chr20 33593260 33593348 CCDS13257.1_cds_0_0_chr20_33593261_f 0 + CCDS13257.1_cds_0_0_chr20_33593261_fchr20+ +chr21 32707032 32707192 CCDS13614.1_cds_0_0_chr21_32707033_f 0 + CCDS13614.1_cds_0_0_chr21_32707033_fchr21+ +chr21 32869641 32870022 CCDS13615.1_cds_0_0_chr21_32869642_r 0 - CCDS13615.1_cds_0_0_chr21_32869642_rchr21- +chr21 33321040 33322012 CCDS13620.1_cds_0_0_chr21_33321041_f 0 + CCDS13620.1_cds_0_0_chr21_33321041_fchr21+ +chr21 33744994 33745040 CCDS13625.1_cds_0_0_chr21_33744995_r 0 - CCDS13625.1_cds_0_0_chr21_33744995_rchr21- +chr22 30120223 30120265 CCDS13897.1_cds_0_0_chr22_30120224_f 0 + CCDS13897.1_cds_0_0_chr22_30120224_fchr22+ +chr22 30160419 30160661 CCDS13898.1_cds_0_0_chr22_30160420_r 0 - CCDS13898.1_cds_0_0_chr22_30160420_rchr22- +chr22 30665273 30665360 CCDS13901.1_cds_0_0_chr22_30665274_f 0 + CCDS13901.1_cds_0_0_chr22_30665274_fchr22+ +chr22 30939054 30939266 CCDS13903.1_cds_0_0_chr22_30939055_r 0 - CCDS13903.1_cds_0_0_chr22_30939055_rchr22- +chr5 131424298 131424460 CCDS4149.1_cds_0_0_chr5_131424299_f 0 + CCDS4149.1_cds_0_0_chr5_131424299_fchr5+ +chr5 131556601 131556672 CCDS4151.1_cds_0_0_chr5_131556602_r 0 - CCDS4151.1_cds_0_0_chr5_131556602_rchr5- +chr5 131621326 131621419 CCDS4152.1_cds_0_0_chr5_131621327_f 0 + CCDS4152.1_cds_0_0_chr5_131621327_fchr5+ +chr5 131847541 131847666 CCDS4155.1_cds_0_0_chr5_131847542_r 0 - CCDS4155.1_cds_0_0_chr5_131847542_rchr5- +chr6 108299600 108299744 CCDS5061.1_cds_0_0_chr6_108299601_r 0 - CCDS5061.1_cds_0_0_chr6_108299601_rchr6- +chr6 108594662 108594687 CCDS5063.1_cds_0_0_chr6_108594663_f 0 + CCDS5063.1_cds_0_0_chr6_108594663_fchr6+ +chr6 108640045 108640151 CCDS5064.1_cds_0_0_chr6_108640046_r 0 - CCDS5064.1_cds_0_0_chr6_108640046_rchr6- +chr6 108722976 108723115 CCDS5067.1_cds_0_0_chr6_108722977_f 0 + CCDS5067.1_cds_0_0_chr6_108722977_fchr6+ +chr7 113660517 113660685 CCDS5760.1_cds_0_0_chr7_113660518_f 0 + CCDS5760.1_cds_0_0_chr7_113660518_fchr7+ +chr7 116512159 116512389 CCDS5771.1_cds_0_0_chr7_116512160_r 0 - CCDS5771.1_cds_0_0_chr7_116512160_rchr7- +chr7 116714099 116714152 CCDS5773.1_cds_0_0_chr7_116714100_f 0 + CCDS5773.1_cds_0_0_chr7_116714100_fchr7+ +chr7 116945541 116945787 CCDS5774.1_cds_0_0_chr7_116945542_r 0 - CCDS5774.1_cds_0_0_chr7_116945542_rchr7- +chr8 118881131 118881317 CCDS6324.1_cds_0_0_chr8_118881132_r 0 - CCDS6324.1_cds_0_0_chr8_118881132_rchr8- +chr9 128764156 128764189 CCDS6914.1_cds_0_0_chr9_128764157_f 0 + CCDS6914.1_cds_0_0_chr9_128764157_fchr9+ +chr9 128787519 128789136 CCDS6915.1_cds_0_0_chr9_128787520_r 0 - CCDS6915.1_cds_0_0_chr9_128787520_rchr9- +chr9 128882427 128882523 CCDS6917.1_cds_0_0_chr9_128882428_f 0 + CCDS6917.1_cds_0_0_chr9_128882428_fchr9+ +chr9 128937229 128937445 CCDS6919.1_cds_0_0_chr9_128937230_r 0 - CCDS6919.1_cds_0_0_chr9_128937230_rchr9- +chrX 122745047 122745924 CCDS14606.1_cds_0_0_chrX_122745048_f 0 + CCDS14606.1_cds_0_0_chrX_122745048_fchrX+ +chrX 152648964 152649196 CCDS14733.1_cds_0_0_chrX_152648965_r 0 - CCDS14733.1_cds_0_0_chrX_152648965_rchrX- +chrX 152691446 152691471 CCDS14735.1_cds_0_0_chrX_152691447_f 0 + CCDS14735.1_cds_0_0_chrX_152691447_fchrX+ +chrX 152694029 152694263 CCDS14736.1_cds_0_0_chrX_152694030_r 0 - CCDS14736.1_cds_0_0_chrX_152694030_rchrX-