After testing your code and data, I can see Maple hitting a couple of snags with import and joining of TimeSeries data. For reasons that I'm not yet clear on, the import of your data from the Excel xlsx comes in as an integer value that Maple doesn't play well with. To get around this, I split your data into two separate csv files and read those in.
After doing so using ImportMatrix, I didn't have any immediate success merging the TimeSeries data, so I instead tried using a DataFrame, which is slightly more agnostic to incoming data. I usually use DataFrames whenever I have row or column labels, which in this case are the dates and column headers respectively. Storing the data in a DataFrame also makes it much easier to apply any relevant Statistics commands to the data. You might still be able to import your data using ExcelTools:-Import and other techniques, but this seemed like the quickest approach to me.
By default, the Import command uses a DataFrame for storage:
PMEAS_ROUGH := Import( "this:///PM.csv" );
QGAS_ROUGH := Import( "this:///QG.csv" );
The Append command merges the two DataFrames, however the row labels are not sorted.
DF := Append( PMEAS_ROUGH, QGAS_ROUGH );
Reorder the DataFrame, sorting on the row labels:
DF := DF[ sort(RowLabels(DF)), ..];
Optional: Convert the resulting DataFrame to a TimeSeries:
TS := TimeSeriesAnalysis:-TimeSeries( Matrix( < < RowLabels( DF ) > | convert( DF, Matrix )> ), headers = 1 , dates = 1);
I've added an example worksheet (in a zipped Workbook with attached data csv files) - DataFrame_Example.zip
Hope this helps.