Difference between revisions of "Talk:Test includematch"

Revision as of 22:00, 15 December 2007

Includematch searches all fields instead of just one (bug?)

I tested the following example and noticed strange includematch behaviour.

<dpl>
category=African Country
includepage={African Country}:languages,{African Country}:goverment,{African Country}:area,{African Country}:population,{African Country}:gdp,{African Country}:independence_day
mode=userformat
listseparators=¶{|class=sortable ¶!No ¶! Name ¶! Languages ¶! Government ¶! Area ¶! Population ¶! GDP ¶! Independence Day , ¶|-¶|%NR%¶|[[%PAGE%]] , ,¶|}
secseparators=¶|
resultsheader=Number of data = %PAGES%
includematch=/languages\s*=\s*.*English.*/s
</dpl>

Try for example replacing

includematch=/languages\s*=\s*.*English.*/s

with

includematch=/languages\s*=\s*.*1961.*/s

and you still got a result because the "Independance day" column is searched. Shouldn't only the "Languages" column being searched? --Mark P. 23:25, 22 November 2007 (CET)

Reply and Explanation

First, the DPL statement above could be written more elegantly, using the table command. The table command was invented recently to facilitate standard view definitions.

Second the regexep uses ".*" before 'English' or '1961'. This means that also pipe characters will be skipped - so '1961' will be accepted if it occurs ANYWHERE AFTER 'languages'. What you want, however, is to stop the pattern matching at field boundaries /(i.e. at pipoe characters). Therefore you must use [^|]* instead of .* which will match all characters EXCEPT The field delimiting pipe character.

Third, the '.*' AFTER 'English' is useless and should be left away.

<dpl>
 category=African Country
 namespace=
 include={African Country}:languages:goverment:area:population:gdp:independence_day
 includematch=/languages\s*=\s*[^|]*English/s
 table=class=sortable,Name,Languages,Government,Area,Population,GDP,Independence Day
 resultsheader=Number of data = %PAGES%\n
 noresultsheader= no matching country found.\n
</dpl>

And here is the result with 'French':

Number of data = 3

.. and with '1961'

Number of data = 1

Note: If you use the #dpl parser function notation you must not use the pipe directly because it would be misperceived as a field demiter for #dpl itself. In that case you should use something like [^{{!}}]* (assuming that there is a template called ! which returns a single pipe character.

Gero 12:13, 24 November 2007 (CET)

Thank you for this clear explanation. --Mark P. 22:35, 24 November 2007 (CET)

Similar Application - need direction

attempt on the Template:Person page

1. Person is a template that has a parameter Name(which is set to the Person's name- simple right?)
2. Servers is a template that has a parameter Admin1 (which is hard coded to a Person's name)
3. We are attempting to query Servers (of Template:Servers) with Admin1=Name (the person's name from {Person}:Name)
4. this dpl tag set is located in Template:Person.

<DPL>
 category=Server|Person
 notnamespace=Template
 include={Servers}:Status
 includematch=/Admin1\s*=\s*[^|]*{{{Name}}}/s
 table=class="sortable server_attribs",Name,Status
 resultsheader=Number of data = %PAGES%\n
 noresultsheader= No servers administered.\n
</DPL>

attempt on shmo's page

1. shmo is a Person (Template:Person)
2. attempting to use a harded coded value of "shmo" to match to {Servers}:Admin1 (shmo can be hard coded here since it's his page and no one elses)

{{#dpl:category=Server|includematch=/{Servers}:Admin1\s*=\s*shmo/s|notnamespace=Template}}