I was wondering...
How about being able to search multiple tables using the same query in one shot?
This would avoid doing repetitive work since currently one has to create the same search query per table.
There are a couple of utilities that allow search sd files without having to upload the sd file to a data base platform.
Something similar in Instant JChem would be great to have.
> How about being able to search multiple tables using the same query in one shot?
> This would avoid doing repetitive work since currently one has to create the same search query per table.
This is a nice idea, but actually quite complex. Different tables might have different fields (columns) and data types. There would need to be some complex mapping between the query and the different tables. We have thought of providing a "virtual table"that wrapped multiple individual tables, and allowing searches to be parallelised.
This could be quite powerful as it would allow generation of a composite database (e.g. something like ACD with a database table for each vendor) and allow each table to be updated independently, and would allow search to be parallelised if each database if on a different server.
But there is a lot of complexity involved here, so I don't think this is something that is likley to be developed for some time, (unless there is stong demand!)
You can of course import multiple SD files into the same JChem table to achieve this.
One useful trick is to create a field to identify the source file, and set that to a sensible value using a default value for that field when importing. That way you can easily identify the source of each structure, even though you have combined them into a single database table.
>There are a couple of utilities that allow search sd files without having to upload the sd file to a data base platform.
> Something similar in Instant JChem would be great to have.
I don't think this is a route we want to go down. The real power of IJC is that it moves you away from "dead" SD files, and provides proper data typing (not all fields as text) and allows fast and flexible searching.
A key paradigm of IJC is to make it easy to load your SD file into a database where it can be searched properly.
For that reason we are only providing limited direct support for files. Working with your data within a database is so much better.
All your comments are very good valid points.
But... Is there any reason why multiple table searching can't be done at the minimum common-denominator level, in this case molecule searching?
This would eliminate the need to "standardize" and merge multiple sd files into 1 single DB/table.
What do you think?
We'll look at how this can be done.
Thanks for the suggestion!