Our Foxpro app has 2 directories of dbf files (2 companies, each have a different set of dbf's in their own respective directories). However both companies do share a couple dbfs located in "Company A's" directory. We had been working with separate datasource A and datasource B. On the couple of occasions datasource B needed to join a table from datasource A, we had created some "symbolic links" to the shared DBF and IDX files so that our ColdFusion app could access the proper tables. This works, though seems less than ideal.
Today we discovered "catalogs" in Hxtt DBF driver which allows us to properly join tables in separate directories by creating one "parent directory" datasource and fully qualifying the path to those tables in every query. This is certainly more verbose, but MUCH more flexible and is how we are accustomed to working in SQL databases.
However, it seems that there is a performance hit when having one giant parent datasource with multiple subdirectories over the symlink option?
Is there a reason for this? Would it actually be better to go back to the symlink option? Are there pros/cons to either approach?
|
|
You can use catalogs, which hasn't performance issue, if you want to do join query on two subdirectories.
|
Main
webmaster@hxtt.com