Bernhard S
Member³
Since I often need to compare schemas, I find it annoying, how long "Compare User Objects" takes. I have about 500 user objects in both, SRC (=source) and TGT (=target) schema. This takes about 5 minutes on my PC!
I have the following ideas about improving usability and speed of such comparisons:
1. To do a complete compare between SRCand TGTschema, it's currently necessary to do a compare both ways: SRC -> TGT and TGT -> SRC
If you just do the first, you won't get shown any difference regarting objects that only exist on TGT and vice versa. This takes either double the time, if you do it sequentially or further slows down comparisons, if you do in parallel.
In order to avoid that I'd suggest a mode which looks at and compares the FULL OUTER JOIN equivalent of both, SRC and TGT objects, not just the LEFT OUTER JOIN equivalent.
2. As I often don't know before looking at the comparison results and in order to speed up the total adaption time, I'd suggest to separate the comparison process from the adaption process. In a first step differences should only be marked on object level and in a second step it should be possible to choose between adapting SRC to TGT or the other way round.
3. When you do comparisons regularly, it shouldn't be necessary to compare all objects of a schema in order to get all the differences. So one way to improve comparison speed might be to only consider objects that have changed in the last [n] days, where [n] should be adjustable by the user. It would help a lot, if PSD could keep track of differences and adaption dates and intelligently use this information to determine a reasonable value for [n].
4. I often need to compare more than two schemas. So a mode to compare a SRC with several TGTs at once would help a lot as well. Since it's always the same SRC and TGTs, it would help, if such comparison settings including the options could be stored and easily reloaded.
Any other ideas to improve performance and usability of "Compare User Objects" are very much appreciated.
I have the following ideas about improving usability and speed of such comparisons:
1. To do a complete compare between SRCand TGTschema, it's currently necessary to do a compare both ways: SRC -> TGT and TGT -> SRC
If you just do the first, you won't get shown any difference regarting objects that only exist on TGT and vice versa. This takes either double the time, if you do it sequentially or further slows down comparisons, if you do in parallel.
In order to avoid that I'd suggest a mode which looks at and compares the FULL OUTER JOIN equivalent of both, SRC and TGT objects, not just the LEFT OUTER JOIN equivalent.
2. As I often don't know before looking at the comparison results and in order to speed up the total adaption time, I'd suggest to separate the comparison process from the adaption process. In a first step differences should only be marked on object level and in a second step it should be possible to choose between adapting SRC to TGT or the other way round.
3. When you do comparisons regularly, it shouldn't be necessary to compare all objects of a schema in order to get all the differences. So one way to improve comparison speed might be to only consider objects that have changed in the last [n] days, where [n] should be adjustable by the user. It would help a lot, if PSD could keep track of differences and adaption dates and intelligently use this information to determine a reasonable value for [n].
4. I often need to compare more than two schemas. So a mode to compare a SRC with several TGTs at once would help a lot as well. Since it's always the same SRC and TGTs, it would help, if such comparison settings including the options could be stored and easily reloaded.
Any other ideas to improve performance and usability of "Compare User Objects" are very much appreciated.