Hi Marco,
There is a truncation based on file names. I've constructed an
example similar to what you described, and this is what I see:
yes? use
testfile9ABCDEF0123456789ABCDEF0123456789ABCDEF0.nc
yes? use testfile9ABCDEF0123456789ABCDEF0.nc
yes? use testfile9ABCDEF012345678.nc
yes? use testfile9ABCDEF0123456789A.nc
yes? use testfile9ABCDEF0123456789AB.nc
yes? use testfile9ABCDEF0123456789ABC.nc
yes? use
testfile9ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0.nc
yes? sh data testfile9ABCDEF0123456789A.nc
currently SET data sets:
4> ./sub/testfile9ABCDEF0123456789A.nc
name title I J K
L M N
AA {1,2,3,4,5} 1:5 ... ...
... ... ...
yes? sh dat
testfile9ABCDEF0123456789ABCDEF0123456789ABCDEF0.nc
currently SET data sets:
2> ./sub/testfile9ABCDEF0123456789ABCDEF0.nc
name title I
J K L M N
AA {1,2,3,4,5} 1:5 ...
... ... ... ...
I'm seeing the result for testfile9ABCDEF0123456789A.nc , but in the
second "show" command, the name given in the "show data" command is
truncated at 32 characters, so it matches the second file name in
the list. That is what is happening in your later example.
I'm not sure why your example isn't returning a result for "sh
data testfile9ABCDEF0123456789A.nc", but we'll certainly
give the file name comparison more working room. The length of that
string comparison for dataset names has been the same, going back at
least 10 years, so it has not changed recently.
On 9/29/2015 9:23 AM, Marco van Hulten
wrote:
Hi fellow Ferretters,
Sometimes loaded datasets are not recognised by filename, and I get the
error message "unknown data set". I am not able to make a minimal
working example that consistently reproduces this result. It might be
related to files being in subdirectories and/or their names being
longer than a certain length, but I cannot be certain.
I first make a test file:
let aa = {1,2,3,4,5}
save/file=testfile aa
After that I exit Ferret, make a directory "sub", adjust my FER_DATA to
contain only . and sub, and put the file in there. Furthermore, I make
some symlinks of different filename lengths. I restart Ferret and
load several of the files (which are the same in the end):
yes? sh data testfile9ABCDEF0123456789ABC.nc
**ERROR: unknown data set:
yes? use testfile9ABCDEF0123456789A.nc
yes? sh data testfile9ABCDEF0123456789A.nc
currently SET data sets:
4> sub/testfile9ABCDEF0123456789A.nc (default)
name title I J K L M N
AA {1,2,3,4,5} 1:5 ... ... ... ... ...
yes? use testfile9ABCDEF0123456789ABC.nc
yes? sh data testfile9ABCDEF0123456789ABC.nc
currently SET data sets:
6> sub/testfile9ABCDEF0123456789ABC.nc (default)
name title I J K L M N
AA {1,2,3,4,5} 1:5 ... ... ... ... ...
yes?
First I cannot approach testfile9ABCDEF0123456789ABC.nc, while after
opening/showing another one I *can*. How could this be happening?
Just another quick observation, probably a side effect of allowing
short-hand writing in Ferret:
yes? sh data testfile9ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0.nc
currently SET data sets:
2> sub/testfile9ABCDEF0123456789ABCDEF0.nc
name title I J K L M N
AA {1,2,3,4,5} 1:5 ... ... ... ... ...
yes? sh data/brief
currently SET data sets:
1> sub/testfile9ABCDEF0123456789ABCDEF0123456789ABCDEF0.nc
2> sub/testfile9ABCDEF0123456789ABCDEF0.nc
3> sub/testfile9ABCDEF012345678.nc
4> sub/testfile9ABCDEF0123456789A.nc
5> sub/testfile9ABCDEF0123456789AB.nc
6> sub/testfile9ABCDEF0123456789ABC.nc
7> sub/testfile9ABCDEF0123456789ABCDEF0123456789ABCDEF0123456789ABCDEF0.nc (default)
Marco
|