String gets mysteriously cut off

Question

In my application I use WpfLocalization to provide translations while the application is running. The library will basically maintain a list of properties and their assigned localization keywords and use DependencyObject.SetValue() to update their values when the active language is changed.

The scenario in which I noticed my problem is this: I have a simple TextBlock and have assigned a localization keyword for its Text property. Now when my application starts, it will write the initial value into it and it will display just fine on screen. Now I switch the language and the new value is set as the Text property but only half the text will actually display on screen. Switching the languages back and forth does not have any effect. The first language is always displayed fine, the second is cut off (in the middle of words, but always full characters).

The relative length of both languages to each other does not seem to have anything to do with it. In my test case the working language string is 498 bytes and the one that gets cut off is 439 bytes and gets cut off after 257 bytes).

When I inspect the current value of the Text property of said TextBlock right before I change its value through the localization code, it will always have the expected value (not cut off) in either language.

When inspecting the TextBlock at runtime through WPF Inspector it will display the cut off text as the Text property in the second language.

This makes no sense to me at all thus far. But now it gets better.

The original WpfLocalization library reads the localized strings from standard resource files, but we use a modified version that can also read those string from an Excel file. It does that by opening an OleDbConnection using the Microsoft OLE DB driver and reading the strings through that. In the debugger I can see that all the values are read just fine.

Now I was really surprised when a colleague found the fix for the "cut off text" issue. He re-ordered the rows in the Excel sheet. I don't see how that could be relevant, but switching between the two versions of that file has an impact on the issue.

Answer 1

That does actually make sense, it's because the ole db driver for Excel has to take a sample of the data in a column to assign it a type and in the case of string, also a length. If it only samples values below the 255 character threshold, you will get a string(255) type and truncated text, if it has sampled a longer string, it will assign it as a memo column and allow longer strings to be retrieved / stored. By re-ordering, you are changing which rows are sampled.

If you read the SQL Server to Excel using oledb you will find this is a known issue. http://msdn.microsoft.com/en-us/library/ms141683.aspx - since you are using the same ole db driver, I would expect the situation to also apply to you.

From the docs:

Truncated text. When the driver determines that an Excel column contains text data, the driver selects the data type (string or memo) based on the longest value that it samples. If the driver does not discover any values longer than 255 characters in the rows that it samples, it treats the column as a 255-character string column instead of a memo column. Therefore, values longer than 255 characters may be truncated. To import data from a memo column without truncation, you must make sure that the memo column in at least one of the sampled rows contains a value longer than 255 characters, or you must increase the number of rows sampled by the driver to include such a row. You can increase the number of rows sampled by increasing the value of TypeGuessRows under the HKEY_LOCAL_MACHINE\\SOFTWARE\\Microsoft\\Jet\\4.0\\Engines\\Excel registry key. For more information, see PRB: Transfer of Data from Jet 4.0 OLEDB Source Fails w/ Error.