|
4
|
1 <tool id="tp_easyjoin_tool" name="Join" version="@BASE_VERSION@.0">
|
|
0
|
2 <description>two files</description>
|
|
4
|
3 <macros>
|
|
|
4 <import>macros.xml</import>
|
|
|
5 </macros>
|
|
|
6 <expand macro="requirements">
|
|
|
7 <requirement type="set_environment">TP_SCRIPT_PATH</requirement>
|
|
|
8 </expand>
|
|
|
9 <version_command>join --version | head -n 1</version_command>
|
|
|
10 <command>
|
|
|
11 <![CDATA[
|
|
|
12 cp \$TP_SCRIPT_PATH/sort-header ./ &&
|
|
|
13 chmod +x sort-header &&
|
|
|
14 perl \$TP_SCRIPT_PATH/easyjoin
|
|
|
15 $jointype
|
|
|
16 -t ' '
|
|
|
17 $header
|
|
|
18 -e '$empty_string_filler'
|
|
|
19 -o auto
|
|
|
20 $ignore_case
|
|
|
21 -1 '$column1'
|
|
|
22 -2 '$column2'
|
|
|
23 "$infile1"
|
|
|
24 "$infile2"
|
|
|
25 > '$output'
|
|
|
26 ]]>
|
|
0
|
27 </command>
|
|
4
|
28 <inputs>
|
|
|
29 <param name="infile1" format="tabular" type="data" label="1st file" />
|
|
|
30 <param name="column1" label="Column to use from 1st file" type="data_column" data_ref="infile1" accept_default="true" />
|
|
0
|
31
|
|
4
|
32 <param name="infile2" format="txt" type="data" label="2nd File" />
|
|
|
33 <param name="column2" label="Column to use from 2nd file" type="data_column" data_ref="infile2" accept_default="true" />
|
|
0
|
34
|
|
|
35 <param name="jointype" type="select" label="Output lines appearing in">
|
|
4
|
36 <option value=" " selected="True">Both 1st & 2nd file.</option>
|
|
|
37 <option value="-v 1">1st but not in 2nd file. (-v 1)</option>
|
|
|
38 <option value="-v 2">2nd but not in 1st file. (-v 2)</option>
|
|
|
39 <option value="-a 1">Both 1st & 2nd file, plus unpairable lines from 1st file. (-a 1)</option>
|
|
|
40 <option value="-a 2">Both 1st & 2nd file, plus unpairable lines from 2st file. (-a 2)</option>
|
|
|
41 <option value="-a 1 -a 2">All lines [-a 1 -a 2]</option>
|
|
0
|
42 <option value="-v 1 -v 2">All unpairable lines [-v 1 -v 2]</option>
|
|
|
43 </param>
|
|
|
44
|
|
4
|
45 <param name="header" type="boolean" checked="false" truevalue="--header" falsevalue=""
|
|
|
46 label="First line is a header line" help="Use if first line contains column headers. It will not be sorted." />
|
|
|
47 <param name="ignore_case" type="boolean" checked="false" truevalue="-i" falsevalue=""
|
|
|
48 label="Ignore case" help="Sort and Join key column values regardless of upper/lower case letters." />
|
|
0
|
49 <param name="empty_string_filler" type="text" size="20" value="0" label="Value to put in unpaired (empty) fields">
|
|
|
50 <sanitizer>
|
|
|
51 <valid initial="string.printable">
|
|
|
52 <remove value="'"/>
|
|
|
53 </valid>
|
|
|
54 </sanitizer>
|
|
|
55 </param>
|
|
4
|
56 </inputs>
|
|
|
57 <outputs>
|
|
|
58 <data name="output" format="input" metadata_source="infile1"/>
|
|
|
59 </outputs>
|
|
|
60 <tests>
|
|
|
61 <test>
|
|
|
62 <param name="infile1" value="easyjoin1.tabular" />
|
|
|
63 <param name="column1" value="1" />
|
|
|
64 <param name="infile2" value="easyjoin2.tabular" />
|
|
|
65 <param name="column2" value="1" />
|
|
|
66 <param name="header" value="True" />
|
|
|
67 <param name="jointype" value="-a 1 -a 2" />
|
|
|
68 <output name="output" file="easyjoin_result1.tabular" />
|
|
|
69 </test>
|
|
|
70 </tests>
|
|
|
71 <help>
|
|
|
72 <![CDATA[
|
|
0
|
73 **What it does**
|
|
|
74
|
|
|
75 This tool joins two tabular files based on a common key column.
|
|
|
76
|
|
|
77 -----
|
|
|
78
|
|
|
79 **Example**
|
|
|
80
|
|
|
81 **First file**::
|
|
|
82
|
|
|
83 Fruit Color
|
|
|
84 Apple red
|
|
|
85 Banana yellow
|
|
|
86 Orange orange
|
|
|
87 Melon green
|
|
|
88
|
|
|
89 **Second File**::
|
|
|
90
|
|
|
91 Fruit Price
|
|
|
92 Orange 7
|
|
|
93 Avocado 8
|
|
|
94 Apple 4
|
|
|
95 Banana 3
|
|
|
96
|
|
|
97 **Joining** both files, using **key column 1** and a **header line**, will return::
|
|
|
98
|
|
|
99 Fruit Color Price
|
|
|
100 Apple red 4
|
|
|
101 Avocado . 8
|
|
|
102 Banana yellow 3
|
|
|
103 Melon green .
|
|
|
104 Orange orange 7
|
|
|
105
|
|
4
|
106 .. class:: infomark
|
|
0
|
107
|
|
4
|
108 * Input files need not be sorted.
|
|
|
109 * The header line (**Fruit Color Price**) was joined and kept as first line.
|
|
|
110 * Missing values ( Avocado's color, missing from the first file ) are replaced with a period character.
|
|
0
|
111
|
|
4
|
112 @REFERENCES@
|
|
|
113 ]]>
|
|
|
114 </help>
|
|
0
|
115 </tool>
|
