On 24/06/13 16:01, H.Merijn Brand wrote:
In the git branch "f_dir-ref" (a wrong name now, as we changed some
things in the development process), you can find a new extension
DBD::File now understands a new attribute f_dir_search:
f_dir_search
This optional attribute can be set to pass a list of folders to also
find existing tables. It will not be used to create new files.
That does not make sense to me. Does it really mean:
This optional attribute can be set to a list of folders which will be searched
for existing tables and is not used when creating new tables.
BTW, I presume you normally use the word folder (bit of a Windows thing) rather
than directory (f_dir_search also suggests directory not folder).
f_dir_search => [ "/data/bar/csv", "/dump/blargh/data" ],
A real-world example might explain this in more detail using DBD::CSV
my $dbh = DBI->connect ("dbi:CSV:", undef, undef, {
f_schema => undef,
f_dir => "tmp",
f_dir_search => [ "sandbox", "/tmp" ],
f_ext => ".csv/r",
f_lock => 2,
f_encoding => "utf8",
RaiseError => 1,
PrintError => 1,
FetchHashKeyName => "NAME_lc",
}) or die "$DBI::errstr\n";
This will set the default folder for CSV files to be located in the
folder "tmp". This folder will also be used for creating tables.
When opening a handle to an already existing table, and that table's
file(s) cannot be found in the folder indicated by "f_dir", the folders
in the "f_dir"search" (if present) will be searched for the files
needed to open the table.
my $sth = $dbh->prepare ("select * from foo");
will, using the above example, search for
tmp/foo.csv
sandbox/foo.csv
/tmp/foo.csv
The DBI test suite tests this option in t/51dbm_file.t
If people have feedback, it would be appreciated. Both possitive and
negative. If no objections exist, I'd like to merge this branch into
master and request a new release, so we can release DBD::CSV shortly
after
Martin
