Avoiding extra delays when adding multiple gedcoms or merging family files
Tip version: 13.2
Revision date: 13 May 2001
Software version: v5.2-v8
This one won't be for everyone or be used very often.
One feature of Easytree which has changed considerably in behavior between the older software (v 4.2) and the newer software (v 5.2+ is how indexes are built.
The old software gave you the option of either rebuilding the index automatically when you had added individuals and then later called the index, or doing it manually by pressing the update index button from the index screen. Note that the index only got rebuilt on automatic when you accessed that screen or called on a function that accessed that data. These options were offered as rebuilding the index took some time.
Generations v4.2 name index window. Note that "Update Automatically" is checked.
The new software (v5.2 and above) takes a different approach. When a file is first built (upgraded from an older v4.2 family file or imported via gedcom) the index is built at that time. This can take some considerable amount of time with a large file. Afterwards the index is updated "on the fly" as you type in new data or revise the names in your database. This results in a considerable increase in speed during normal operations.
One complication that arises is that you encounter the initial, time-consuming index build whenever you merge two family files or import a gedcom file. The time to build an index is a nonlinear function of the size of the family file. Small files have their indices built much faster than large files.
So, in the event that you have a number of gedcom files to import into your main family file or Generations family files you want to merge, it saves a considerable amount of time to import/merge the small files together into a working family file. Then you merge the working file into your main family file. This way the most time consuming index build (of the largest file), is done only once instead of every time you add another small file.
REMEMBER TO MAKE BACKUPS BEFORE YOU DO THIS, JUST IN CASE SOMETHING GOES WRONG OR YOU DECIDE THAT YOU MADE A MISTAKE!
Also, you will probably want to use the "Automatic" source option to indicate the different original files that you started with - it is MUCH easier to use this to document where the data originated as you import it than to backtrack later if you realize you need to improve your documentation.
Changes with the v8.5/8.5.1 release
The problem of the large time to sort or resort the name index was addressed with the v8.5 release and the v8.5.1 patch as applied to the v8 releases. There was a major improvment in the performance of the sort, yielding index sorts 8-12x (yes, times, not percent) faster than with the earlier releases. The earlier releases had also required a considerable amount of disk space for temporary files during the index build/rebuild. This has been largely eliminated with the new sorting scheme.
Regards, Mike Hobart
Copyright © 1999-2001 Michael A. Hobart, commercial reuse restricted.