| 1 | == History Table == |
| 2 | |
| 3 | === proc_cmd column === |
| 4 | |
| 5 | This column records the processing command that was used. In the case of pvoconvert, it records the command-line statement that produced the fits file. For instance "pvoconvert filename" for simple conversion of a single file. |
| 6 | |
| 7 | To facilitate conversion of lots of the survey files, pvoconvert has a "pvoconvert -D directory" option, where you can just give a directory and all files therein will be converted. This has been created as an alternative to using "pvoconvert directory/*", as the latter was writing the following to the proc_cmd column in the history table: "pvoconvert directory/file1 directory/file2 directory/file3 ..." and so on. This rapidly filled up the space allocated in that table column. |
| 8 | |
| 9 | This raises the question of what information do we want recorded in the proc_cmd column. Is it more important to show the actual command as typed, or to show the file where the data in the FITS file came from? [[BR]] |
| 10 | If the former, then the use of the directory option above will result in proc_cmd having "pvoconvert -D directory" as its value.[[BR]] |
| 11 | If the latter, then it would just show "pvoconvert filename" (where directory/filename has the data for this particular fits file). |
| 12 | |
| 13 | Or, should we just not allow multiple files to be processed at once by pvoconvert, in which case this problem will not arise? |