MultiCol match: empty array means empty filter

let DateTime handle rev column on its own

This fixes filtering for %lastupdate%

a $STRUCT.*$ filter doesn't need reference the same schemas

A filter may reference all schemas of the page the aggregation is on
which are not necessarily the same schemas that are queried.

make DateFormatConverter properly escape stuff

Added a minimal test. More cases would probably be helpful

Initialize DateType with dformat configuration

We're not using strftime in our Date and DateTime types because we want
to be able to represent dates outside the unix timestamp range.

To reuse the s

Initialize DateType with dformat configuration

We're not using strftime in our Date and DateTime types because we want
to be able to represent dates outside the unix timestamp range.

To reuse the strftime format string we have in DokuWiki it needs to be
converted to a date() compatible format. The class added for that could
ultimately be moved to a third party project.

show more ...

properly parse width settings in aggregation table

makes sure only valid CSS values are supplied

allow multicolumn filters #124

when referencing a multi value column through the $STRUCT.<column>$
filter placeholder, all found values are used for the comparison using
an OR sub group.

To make th

allow multicolumn filters #124

when referencing a multi value column through the $STRUCT.<column>$
filter placeholder, all found values are used for the comparison using
an OR sub group.

To make this work some changes had to be introduced. You can only
reference _one_ column in a filter, not multiple. Anything before or
after the column placeholder will be applied to each value in the
multicolumn.

Types overrifding the filter() method need to make sure they handle
arrays in $value correctly.

show more ...

implement a %lastupdated% column #122

make label part of Page column

only allow Page types for PageColumn

Revert "Add $TITLE$ placeholder for aggregations and tests"

This reverts commit 7249dd99c6b513e7a0e270e0532734a1bf838ae4.

As discussed in #133 the $TITLE$ place holder is no longer needed since
tit

Revert "Add $TITLE$ placeholder for aggregations and tests"

This reverts commit 7249dd99c6b513e7a0e270e0532734a1bf838ae4.

As discussed in #133 the $TITLE$ place holder is no longer needed since
title columns match against the pid as well. Since gathering the title
data for the filter is actually performance relevant, it makes sense to
remove it again.

show more ...

correctly dynamically filter %title% column #132

use rawValue for old revisions and column filters #131 #129

changed SchemaData to have options for output control

instead of having a whole bunch of paramters to the output functions.
Introduces rawValue cleaning as a new option.

introduce a %title% column for aggregations. fixes #132

Using %title% works like the %pageid% column but sets usetitles to true

run rawValue on found columns only

introduce Type::rawValue(). fixes #131

Add $TITLE$ placeholder for aggregations and tests

add PID to SchemaData select() queries

make use of filter() instead of compare()

isset is not good for testing an array containing NULL

integrate select() into SchemaData

Because the signature of buildGetDataSQL() had to change a lot of tests
are breaking currently. Those tests need to be adjusted first, then we
can see if the imple

integrate select() into SchemaData

Because the signature of buildGetDataSQL() had to change a lot of tests
are breaking currently. Those tests need to be adjusted first, then we
can see if the implemention itself is correct.

After this has been done, the select has to be adjusted to select the
pageid as PID to make it available to possible select() implementations
that rely on the availability of that alias.

show more ...

group by the alias not the raw column

since the type may do somthing more fancy than accessing the column

have types define the column selection

introduce getSelectStatement to modify selects later on