Dan Nelson <dnelson@stripped> writes:
> In the last episode (Jun 02), Jesse F. Hughes said:
>> After a recent hard drive kerfluffle and the results of fsck, I'm left
>> with a slew of jumbled database files. The file command can tell me the
>> file types, like so:
>> #15901614: MySQL table definition file Version 10
>> #15901615: MySQL MISAM compressed data file Version 1
>> #15901617: MySQL table definition file Version 10
>> #15901618: MySQL MISAM compressed data file Version 1
>> #15901620: MySQL table definition file Version 10
>> #15901621: MySQL MISAM compressed data file Version 1
>> These files are, I'm pretty sure, from my mythtv database. I'm
>> rebuilding my box and it would be nice if I can keep the mythtv
>> database the same as before, but how can I tell which file is which?
> "table definition" files are the .frm files; "MISAM compressed data" files
> are .MYI files. Unfortunately, you're missing the .MYD files in that list,
> which don't have a header (sometimes 'file' thinks they're dbase format).
> You may be in luck, though. The filenames are the inode numbers of the
> original files, and it looks like they're in sequential order, so the tables
> were probably created all at once, one after the other. File #15901616 (if
> it exists) is probably the .MYD file that corresponds to #15901614 (.frm)
> and #15901615 (.MYI).
> The numerical order of the files should match the order the mythtv setup
> script creates its tables. If they aren't in order for some reason, then
> you can rename the groups of files into "a.frm", "a.MYI", "a.MYD", "b.frm",
> "b.MYI", "b.MYD", etc, then run "show create table a" etc to determine the
> table layout and match them up to known mythtv tables.
Ouch. I've found the time to look into the problem a little more
I do have a lot of DBase3 files listed, with varying number of
records. Unfortunately, they *aren't* in sequential order with the
.MYI and .frm files. I probably used a MySQL dump file to create
them, so I suppose that the previous files were unlinked and new ones
Except that the dbase files I see have *lower* inode numbers than the
.MYIs and .frms!
Is there some way to match the .MYI to its corresponding .MYD? I have
only 33 to 47 .MYD files (judging from the output of file and their
contiguity), but 74 .MYI files.
Jesse F. Hughes
"This post marks the end of an era in the world of mathematics."
-- James S. Harris and the demise of Galois theory