From: Sven Schöling <s.schoeling@linet-services.de>
Date: Fri, 10 Jun 2016 15:10:05 +0000 (+0200)
Subject: Csv: Doku-Update vor mapping feature
X-Git-Tag: release-3.4.1~51
X-Git-Url: http://wagnertech.de/git?a=commitdiff_plain;h=621248dee40d10dfad3a464d332c2a879f68bd85;p=kivitendo-erp.git

Csv: Doku-Update vor mapping feature
---

diff --git a/SL/Helper/Csv.pm b/SL/Helper/Csv.pm
index 08ed9f44e..50a95faad 100644
--- a/SL/Helper/Csv.pm
+++ b/SL/Helper/Csv.pm
@@ -488,84 +488,105 @@ Examples:
   [ [ 'datatype', 'ordernumber', 'customer', 'transdate' ],
     [ 'datatype', 'partnumber', 'qty', 'sellprice' ] ]
 
-=item C<profile> [{profile => \%ACCESSORS, class => class, row_ident => ri},]
+=item C<profile> PROFILE_DATA
 
-This is an ARRAYREF to HASHREFs which may contain the keys C<profile>, C<class>
-and C<row_ident>.
+The profile mapping csv to the objects.
+
+See section L</PROFILE> for information on this topic.
+
+=item C<ignore_unknown_columns>
+
+If set, the import will ignore unkown header columns. Useful for lazy imports,
+but deactivated by default.
+
+=item C<case_insensitive_header>
+
+If set, header columns will be matched against profile entries case
+insensitive, and on match the profile name will be taken.
+
+Only works if a profile is given, will die otherwise.
+
+If both C<case_insensitive_header> and C<strict_profile> is set, matched header
+columns will be accepted.
+
+=item C<strict_profile>
+
+If set, all columns to be parsed must be specified in C<profile>. Every header
+field not listed there will be treated like an unknown column.
+
+If both C<case_insensitive_header> and C<strict_profile> is set, matched header
+columns will be accepted.
+
+=back
+
+=head1 PROFILE
+
+The profile is needed for mapping csv data to the accessors in the data object.
+
+The basic structure is:
+
+  PROFILE       := [ CLASS_PROFILE, CLASS_PROFILE* ]
+  CLASS_PROFILE := {
+                      profile   => { ACCESSORS },
+                      class     => $classname,
+                      row_ident => $row_ident,
+                   }
+  ACCESSORS     := $field => $accessor, ACCESSORS*
 
 The C<profile> is a HASHREF which may be used to map header fields to custom
 accessors. Example:
 
-  [ {profile => { listprice => listprice_as_number }} ]
+  [
+    {
+      profile => {
+        listprice => 'listprice_as_number',
+      }
+    }
+  ]
 
-In this case C<listprice_as_number> will be used to read in values from the
+In this case C<listprice_as_number> will be used to store the values from the
 C<listprice> column.
 
 In case of a One-To-One relationship these can also be set over
 relationships by separating the steps with a dot (C<.>). This will work:
 
-  [ {profile => { customer => 'customer.name' }} ]
+  customer => 'customer.name',
 
 And will result in something like this:
 
   $obj->customer($obj->meta->relationship('customer')->class->new);
   $obj->customer->name($csv_line->{customer})
 
-But beware, this will not try to look up anything in the database. You will
+Beware, this will not try to look up anything in the database! You will
 simply receive objects that represent what the profile defined. If some of
-these information are unique, and should be connected to preexisting data, you
+these information are unique, or should be connected to preexisting data, you
 will have to do that for yourself. Since you provided the profile, it is
 assumed you know what to do in this case.
 
 If no profile is given, any header field found will be taken as is.
 
 If the path in a profile entry is empty, the field will be subjected to
-C<strict_profile> and C<case_insensitive_header> checking, will be parsed into
-C<get_data>, but will not be attempted to be dispatched into objects.
+C<strict_profile> and C<case_insensitive_header> checking and will be parsed
+into C<get_data>, but will not be attempted to be dispatched into objects.
 
-If C<class> is present, the line will be handed to the new sub of this class,
-and the return value used instead of the line itself.
+C<class> must be present. A new instance will be created for each line before
+dispatching into it.
 
-C<row_ident> is a string to recognize the right profile and class for each data
-line in multiplexed data. It must match the value in the column 'dataype' for
-each class.
+C<row_ident> is used to determine the correct profile in multiplexed data and
+must be given there. It's not used in non-multiplexed data.
 
-In case of multiplexed data, C<class> and C<row_ident> must be given.
 Example:
-  [ {
+  [
+    {
       class     => 'SL::DB::Order',
       row_ident => 'O'
     },
     {
       class     => 'SL::DB::OrderItem',
       row_ident => 'I',
-      profile   => {sellprice => sellprice_as_number}
-    } ]
-
-=item C<ignore_unknown_columns>
-
-If set, the import will ignore unkown header columns. Useful for lazy imports,
-but deactivated by default.
-
-=item C<case_insensitive_header>
-
-If set, header columns will be matched against profile entries case
-insensitive, and on match the profile name will be taken.
-
-Only works if a profile is given, will die otherwise.
-
-If both C<case_insensitive_header> and C<strict_profile> is set, matched header
-columns will be accepted.
-
-=item C<strict_profile>
-
-If set, all columns to be parsed must be specified in C<profile>. Every header
-field not listed there will be treated like an unknown column.
-
-If both C<case_insensitive_header> and C<strict_profile> is set, matched header
-columns will be accepted.
-
-=back
+      profile   => { sellprice => 'sellprice_as_number' }
+    },
+  ]
 
 =head1 ERROR HANDLING