2 added 344 characters in body edited Apr 8 '13 at 16:19 Andy Ross 17.4k22 gold badges5252 silver badges8787 bronze badges This is clearly a bug in part extraction for TemporalData. You can work around it by giving your times explicitly. subset = td["Part", 1, {Cases[td["PathTimes"], x_ /; 10 <= x <= 50]}] I have a feeling this is more in the spirit of what you want here anyway. The "Part" property resamples the data so that it is uniformly spaced if you give {tmin, tmax, dt}. You omit the dt so it has to be inferred. In your case it should be trying to use the smallest time step in your data. This will result in a very large data object in this case because the smallest step is incredibly small compared to the time range you want. EDIT It is additionally worth mentioning that "Part" uses the "PathFunctions" to build the subset. When you just want a window it can be unnecessarily expensive for continuous time process data. It is trivial to get what you want from the data itself and avoid building "PathFunctions". Cases[td["Path", 1], {t_/; 10<=t<=50, _}] This is clearly a bug in part extraction for TemporalData. You can work around it by giving your times explicitly. subset = td["Part", 1, {Cases[td["PathTimes"], x_ /; 10 <= x <= 50]}] I have a feeling this is more in the spirit of what you want here anyway. The "Part" property resamples the data so that it is uniformly spaced if you give {tmin, tmax, dt}. You omit the dt so it has to be inferred. In your case it should be trying to use the smallest time step in your data. This will result in a very large data object in this case because the smallest step is incredibly small compared to the time range you want. This is clearly a bug in part extraction for TemporalData. You can work around it by giving your times explicitly. subset = td["Part", 1, {Cases[td["PathTimes"], x_ /; 10 <= x <= 50]}] I have a feeling this is more in the spirit of what you want here anyway. The "Part" property resamples the data so that it is uniformly spaced if you give {tmin, tmax, dt}. You omit the dt so it has to be inferred. In your case it should be trying to use the smallest time step in your data. This will result in a very large data object in this case because the smallest step is incredibly small compared to the time range you want. EDIT It is additionally worth mentioning that "Part" uses the "PathFunctions" to build the subset. When you just want a window it can be unnecessarily expensive for continuous time process data. It is trivial to get what you want from the data itself and avoid building "PathFunctions". Cases[td["Path", 1], {t_/; 10<=t<=50, _}] 1 answered Apr 8 '13 at 16:03 Andy Ross 17.4k22 gold badges5252 silver badges8787 bronze badges This is clearly a bug in part extraction for TemporalData. You can work around it by giving your times explicitly. subset = td["Part", 1, {Cases[td["PathTimes"], x_ /; 10 <= x <= 50]}] I have a feeling this is more in the spirit of what you want here anyway. The "Part" property resamples the data so that it is uniformly spaced if you give {tmin, tmax, dt}. You omit the dt so it has to be inferred. In your case it should be trying to use the smallest time step in your data. This will result in a very large data object in this case because the smallest step is incredibly small compared to the time range you want.