Good Morning Maik: You are correct, you can not simply rsync in the hgcentral contents and expect your local changes to stay in place. The ideal situation for your local installation is that any new rows you add to tables in hgcentral should be kept as INSERT SQL statements in your own private source tree separate from the operating system. Thus, anytime that hgcentral is updated, you can run those INSERT statements again to add your local organisms.
The SQL statement file we deliver for the contents of hgcentral at: http://hgdownload.cse.ucsc.edu/admin/hgcentral.sql Does have proper update/replace statements for the tables 'clade', 'defaultDb' 'genomeClade', 'hubPublic', and 'targetDb'. For those tables, you would not need to re-run your INSERT statements. *HOWEVER*, for many of the other tables, the entire table is dropped and reloaded freshly from the INSERT statements in the hgcentral.sql file. For those tables you would need to re-run your INSERT SQL statements. --Hiram Maik Richter wrote: > Hello, > > we have added new genome assemblies to our browser mirror using the > automation scripts. This adds entries to the dbDb table in hgcentral (and > possible genomeClade and defaultGenome). > We want to run a periodic rsync to keep up to date with UCSC, however want to > avoid clobbering our changes to these tables. > > As we assume that is a common problem, we wonder if anybody has a good > solution for this? Is there for example a way to merge mysql tables, keeping > the rows that we have added? Sorry if this topic has already been discussed > several times before, but it wasn't stated out clearly enough for us what to > do > in detail. Any help is really appreciated. > > Thanks a lot _______________________________________________ Genome maillist - [email protected] https://lists.soe.ucsc.edu/mailman/listinfo/genome
