# HG changeset patch -- Bitbucket.org # Project galaxy-dist # URL http://bitbucket.org/galaxy/galaxy-dist/overview # User Kelly Vincent <kpvincent@bx.psu.edu> # Date 1282329639 14400 # Node ID 4464d55104b8a33407bf03d822612afd4a679eb0 # Parent d9e099daa8279880d5e054680ab28e8fc3064f77 Updated column join to handle situation when last column is a join column and a row is missing that column --- a/tools/new_operations/column_join.py +++ b/tools/new_operations/column_join.py @@ -208,7 +208,8 @@ def __main__(): hinges = [ h for h in hinges if h ] current, loc = hinges[0], hinge_dict[ hinges[0] ] # first output empty columns for vertical alignment (account for "missing" files) - # write output if trailing empty columns + # write output for leading and trailing empty columns + # columns missing from actual file handled further below current_data = [] if current != old_current: # fill trailing empty columns with appropriate fill value @@ -244,17 +245,25 @@ def __main__(): split_line.append( item ) else: split_line.append( fill_empty[ i + 1 ] ) + # add actual data to be output below if ''.join( split_line ): - # add actual data to be output below for col in cols: if col > hinge: - current_data.append( split_line[ col - 1 ] ) + # if this column doesn't exist, add the appropriate filler or empty column + try: + new_item = split_line[ col - 1 ] + except IndexError: + if fill_empty: + new_item = fill_empty[ col ] + else: + new_item = '' + current_data.append( new_item ) # grab next line for selected file current_lines[ loc ] = tmp_input_files[ loc ].readline().rstrip( '\r\n' ) # write relevant data to file - if current == old_current: + if current == old_current and current_data: fout.write( '%s%s' % ( delimiter, delimiter.join( current_data ) ) ) - else: + elif current_data: fout.write( '%s%s%s' % ( current, delimiter, delimiter.join( current_data ) ) ) last_loc = loc old_current = current --- a/tools/new_operations/column_join.xml +++ b/tools/new_operations/column_join.xml @@ -116,17 +116,29 @@ import simplejson <param name="input" value="column_join_in12.pileup" ftype="pileup" /><output name="output" file="column_join_out4.pileup" ftype="tabular" /></test> - </tests> + <test> + <!-- Test for handling missing column --> + <param name="input1" value="column_join_in13.tabular" ftype="tabular" /> + <param name="hinge" value="1" /> + <param name="columns" value="5" /> + <param name="fill_empty_columns_switch" value="fill_empty" /> + <param name="column_fill_type" value="single_fill_value" /> + <param name="fill_value" value="0" /> + <param name="input2" value="column_join_in14.tabular" ftype="tabular" /> + <param name="input" value="column_join_in15.tabular" ftype="tabular" /> + <output name="output" file="column_join_out5.tabular" ftype="tabular" /> + </test> + </tests><help> **What it does** This tool allows you to join several files with the same column structure into one file, removing certain columns if necessary. The user needs to select a 'hinge', which is the number of left-most columns to match on. They also need to select the columns to include in the join, which should include the hinge columns, too. -Note that the files are expected to have the same number of columns. +Note that the files are expected to have the same number of columns. If for some reason the join column is missing (this only applies to the last column(s)), the tool attempts to handle this situation by inserting an empty item (or the appropriate filler) for that column on that row. This could lead to the situation where a row has a hinge but entirely empty or filled columns, if the hinge exists in at least one file but every file that has it is missing the join column. Also, note that the tool does not distinguish between a file missing the hinge altogether and a file having the hinge but missing the column (in both cases the column would be empty or filled). There is an example of this below. ----- -**Example** +**General Example** Given the following files:: @@ -213,6 +225,36 @@ To join on columns 3 and 4 combining on chr2 83 T 8 chr2 84 A 9 +**Example with missing columns** + +Given the following input files:: + + FILE 1 + 1 A + 2 B b + 4 C c + 5 D + 6 E e + + FILE 2 + 1 M m + 2 N + 3 O o + 4 P p + 5 Q + 7 R r + +if we join only column 3 using column 1 as the hinge and with a fill value of '0', this is what will be output:: + + 1 0 m + 2 b 0 + 3 0 o + 4 c p + 5 0 0 + 6 e 0 + 7 0 r + +Row 5 appears in both files with the missing column, so it's got nothing but fill values in the output file. </help></tool>