转换表或时间表的数据类型

Question

Within a script I am loading data from SQL Server with select and a complex SQL-Statement. The returning datatype seems to be table.

First I am extracting a timestamp column out of this table and do some fancy filtering, to split the dataset into subsets (device is off for some hours, I like to analyse the data set by set). So far, everything is great.

Afterwards I would like to fill the subset every time a dataset is missing. Quite often, the data from a single timestamp is missing, so I would like to insert data with a timetable ('Time'-array is created before):

TT = timetable(Time,data(idTracks(i)+1:idTracks(i+1),{'ID',...
    ...
    'positionLatitude','positionLongitude','speed'}));

retime(TT,'regular','linear','TimeStep',dt) 

I got the error message 'All variables in input timetables must be numeric, datetime, or duration when synchronizing using 'linear'.' I got that but I know, that in the database, all my columns are defined as float and they all contains numbers. I tried to CAST within the SQL-Statement. Does not work. If I try class(data.id) column by column, I alway get a double in return.

How can I easily adjust the datatype, as my timetable has some more columns as mentions in the example (about 30)? I would like to avoid to do the transformation column by column, I would like to fix the table or the timetable in one step if possible, but can't find the trick.