CSViewer: Date values shown differently in same QVD file

[Steeefan]

Hi,

I'm currently working on a QVD generator for a renaming stage. Data is extracted from REST API and stored as QVD files. Those QVD files are then loaded with the files being renamed and some transformations applied, such as creating human-readable date and time fields.

Projects:
NOCONCATENATE LOAD
	msdyn_projectid AS ProjectId,

	Date(Floor(createdon), 'DD.MM.YYYY') AS CreatedOn,
	Date(Floor(modifiedon), 'DD.MM.YYYY') AS ModifiedOn,
	Date(Floor(overriddencreatedon), 'DD.MM.YYYY') AS OverriddenCreatedOn,

	Date(Floor(msdyn_finish), 'DD.MM.YYYY') AS Finish,
	Date(Floor(msdyn_scheduledstart), 'DD.MM.YYYY') AS ScheduledStart,
	Date(Floor(msdyn_taskearlieststart), 'DD.MM.YYYY') AS TaskEarliestStart
FROM
	[lib://01 Extract/MS Planner/Projects.qvd]
	(qvd);

As can be seen, the timestamps are treated all the same. The values are numeric, such as 46106.6153125. In the QVD that is created after the extraction stage, the values shown all in the same format. But not in the QVD that is generated with the code above:

In the Cell Profiler, I'm still seeing the numerical value but different styles for "Formatted value":

I am using the latest greatest version, 2.6.0.4. Why is this happening?

Take care
-Stefan

[vlad_dzhos]

Hi @Steeefan ,

First, I want to make sure I understand the issue correctly. Is the issue specifically with displayed values/formatting, not with the numeric values?

I assume the second screenshot shows a processed file with QVD number+text duals. In that case, during import CSViewer tries to find a supported text representation that matches the date text in the QVD file. It works best on a large enough sample with a close or direct match to a supported format.

In any case, I agree that this is confusing and needs improvement. If possible, please send the processed file (or both) to support@easymorph.com so we can use it to refine the format detection algorithm in future versions.

For now, manually setting the desired display format is the suggested fix.