| Author | Message | ||
     Hakan SIRIN (hsirin) New member Username: hsirin Post Number: 1 Registered: 05-2006 |
Is there anybody out there to tell me how to obtain numerically the values of a history plot to plot it by using another graphic program? | ||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1158 Registered: 06-2003 |
The EXPORT plot modifier works with Histories as well as other plots. See the example "Samples | Misc | Import-Export | Export_History.pde" and the Help entry "Tech Notes | Exporting Data to Other Applications". | ||
| Chris Walton (ccwalton) Member Username: ccwalton Post Number: 9 Registered: 01-2008 |
This may seem a minor point, but when exporting HISTORY data using the EXPORT modifier, in a problem with fixed time steps, is it possible to export the data ONLY on the set time steps? It seems in all the times I've done this FlexPDE records the data on its own chosen time steps in between as well. I can always filter it myself later but it would be a convenience to record only the regular steps from the outset. Thanks.... Chris Walton | ||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1182 Registered: 06-2003 |
Our intention with the History export was to retain as much information as is available in the computation, so as not to miss important behavior by occasional sampling. At present, we have no mechanism for discarding data. Sorry. | ||
| Chris Walton (ccwalton) Member Username: ccwalton Post Number: 10 Registered: 01-2008 |
Understood. But there's still something odd here. I'm simulating temperature evolution in a thick plate subject to a long periodic heat load. I set the simulation to record data every 10s in order to compare to experimental data gathered at that rate: PLOTS for t = 0 by 10 to 3.2e5 ... HISTORY(temp) AT [various position points here] EXPORT If I inspect the data table file shortly after the simulation starts, it contains small time steps determined for each cycle, plus a step at 10n seconds as expected (e.g. 0, 0.599, 1.19...10.00, 10.177...) But when the simulation is finished the file has apparently been written over, and the data at the beginning replaced with much coarser steps, e.g. (125.34, 190.39, 329.99...} no longer including the 10s intervals I asked for. Except the last ~6500 points are back to fine steps like the beginning of the file used to have. Is there some kind of trailing window procedure that winnows the earlier data to keep the file size down? There are ~11000 data points in the final file of ~1MB size, so it would not be very large even if all the data were retained. Thanks, Chris Walton | ||
| Robert G. Nelson (rgnelson) Moderator Username: rgnelson Post Number: 1183 Registered: 06-2003 |
Oops. I forgot about that.... FlexPDE applies a size limit of 16000 points in a history list, after which it applies some compression to the list. I guess this was done on the theory that even a printer page will not have more than a few thousand pixels. This ignores, of course, the tabulation issue. You can override this limit with SELECT HISTORY_LIMIT=100000 I will increase the default on this number in the next maintenance release. |