Forums

There seems to be a discrepency between how data can be
retrieved from a conventional table column and an inserted
column that is added after the DataTable is created.

In a inserted column (created through the
DataTable.insertColumn()
method), a field's value can by
retrieved from the Record object for the field's row by calling
theRecord.getData( fieldName );
where fieldName is the value of the column's "key" attribute.

The value can also be read from the record's _oData array by using
the column's index, a integer, as the array index into the array.
theRecord._oData[ keyInteger ]; // works
but not if the columns key, a string, is used.
theRecord._oData[ indexString ]; // returns undefined

In a conventional column, added through a column configuration
passed to the DataTable constructor, the opposite results are
obtained. Calling getData() or reading _oData[keyIndex] returns
'undefined'; reading _oData[fieldName] will return the correct
value.

The data table's columns appear correct with the inserted column
in the set if ones looks at the configurations returned by
DataTable.getColumnSet().

This matrix describes the results

This behavior seems inconsistent to me. Is it or am I doing
something wrong? It seems reasonable to be abled to access the
data from either type of column in the same way.

A web page demonstration the issue is available at
http://pastebin.com/Qq0NWkzj.
View it in FireFox
using FireBug. Set a breakpoint at the end ShowDataTableValues() to view
data values when the button is clicked. Click the other button to
insert a column.

Thanks in advance for your help.

Thanks for your reply. However, it still doesn't help me understand why retrieving a cell's value in a 'inserted' column should behave differently than a 'configured' column (where the 'configured' column definition is passed to the DataTable constructor.

I understand that in the later case the 'field' attribute of the column is an index into the data array passed via the DataSource (and is interpreted as an integer). This is distinct from the 'key' configuration attribute which is a string, mutually unique among the columns in the data set. (This is a little confusing in the API docs which use 'sField' as the argument to Record;getData() when getData is looking for a 'key'.)

I'm still looking for a way around this difficulty with column types. Practically speaking, I'm looking for a way to iterate over the columns in a table and retrieve their data without having to be aware of the basic nature of the columns.

Thanks again.