A Saturday 22 March 2008, Alex Olivas escrigué:
> > This looks like the NumPy team has chosen for 1.0.5 a different
> > representation for any on its data types. With NumPy 1.0.4 the
> > tests passes to me:
> >
> > $ python tables/tests/test_nestedtypes.py ReadNoReopen.test00a_repr
> > verbose
> > Running ReadNoReopen.test00a_repr
> > ---------------------------------
> > str(tbl)--> /test (Table(2,)) 'test00'
> > repr(tbl)--> /test (Table(2,)) 'test00'
> > description := {
> > "x": Int32Col(shape=(2,), dflt=0, pos=0),
> > "Info": {
> > "value": ComplexCol(itemsize=16, shape=(), dflt=0j, pos=0),
> > "y2": Float64Col(shape=(), dflt=1.0, pos=1),
> > "Info2": {
> > "name": StringCol(itemsize=2, shape=(), dflt='', pos=0),
> > "value": ComplexCol(itemsize=16, shape=(2,), dflt=0j, pos=1),
> > "y3": Time64Col(shape=(2,), dflt=1.0, pos=2),
> > "z3": EnumCol(enum=Enum({'r': 4, 'b': 1, 'g': 2}), dflt='r',
> > base=Int32Atom(shape=(), dflt=0), shape=(2,), pos=3)},
> > "name": StringCol(itemsize=2, shape=(), dflt='', pos=3),
> > "z2": UInt8Col(shape=(), dflt=1, pos=4)},
> > "color": StringCol(itemsize=2, shape=(), dflt=' ', pos=2),
> > "info": {
> > "Name": StringCol(itemsize=2, shape=(), dflt='', pos=0),
> > "Value": ComplexCol(itemsize=16, shape=(), dflt=0j, pos=1)},
> > "y": Float64Col(shape=(2, 2), dflt=1.0, pos=4),
> > "z": UInt8Col(shape=(), dflt=1, pos=5)}
> > byteorder := 'little'
> > chunkshape := (56,)
> >
> > .
> > -------------------------------------------------------------------
> >--- Ran 1 test in 0.087s
> >
> > OK
> >
> > Could you please run the:
> >
> > $ python tables/tests/test_nestedtypes.py ReadNoReopen.test00a_repr
> > verbose
> >
> > line (do not forget the 'verbose' option) and send me back the
> > ouput? I'll update the test to deal with the possible difference in
> > NumPy representations.
> >
> > Thanks,
>
> Here it is:
>
> Running ReadNoReopen.test00a_repr
> ---------------------------------
> str(tbl)--> /test (Table(2L,)) 'test00'
> repr(tbl)--> /test (Table(2L,)) 'test00'
> description := {
> "x": Int32Col(shape=(2,), dflt=0, pos=0),
> "Info": {
> "value": ComplexCol(itemsize=16, shape=(), dflt=(0.0+0.0j),
> pos=0), "y2": Float64Col(shape=(), dflt=1.0, pos=1),
> "Info2": {
> "name": StringCol(itemsize=2, shape=(), dflt='', pos=0),
> "value": ComplexCol(itemsize=16, shape=(2,), dflt=(0.0+0.0j),
> pos=1), "y3": Time64Col(shape=(2,), dflt=1.0, pos=2),
> "z3": EnumCol(enum=Enum({'r': 4, 'b': 1, 'g': 2}), dflt='r',
> base=Int32Atom(shape=(), dflt=0), shape=(2,), pos=3)},
> "name": StringCol(itemsize=2, shape=(), dflt='', pos=3),
> "z2": UInt8Col(shape=(), dflt=1, pos=4)},
> "color": StringCol(itemsize=2, shape=(), dflt=' ', pos=2),
> "info": {
> "Name": StringCol(itemsize=2, shape=(), dflt='', pos=0),
> "Value": ComplexCol(itemsize=16, shape=(), dflt=(0.0+0.0j),
> pos=1)}, "y": Float64Col(shape=(2, 2), dflt=1.0, pos=4),
> "z": UInt8Col(shape=(), dflt=1, pos=5)}
> byteorder := 'little'
> chunkshape := (56,)
> /test (Table(2L,)) 'test00'
> description := {
> "x": Int32Col(shape=(2,), dflt=0, pos=0),
> "Info": {
> "value": ComplexCol(itemsize=16, shape=(), dflt=(0.0+0.0j),
> pos=0), "y2": Float64Col(shape=(), dflt=1.0, pos=1),
> "Info2": {
> "name": StringCol(itemsize=2, shape=(), dflt='', pos=0),
> "value": ComplexCol(itemsize=16, shape=(2,), dflt=(0.0+0.0j),
> pos=1), "y3": Time64Col(shape=(2,), dflt=1.0, pos=2),
> "z3": EnumCol(enum=Enum({'r': 4, 'b': 1, 'g': 2}), dflt='r',
> base=Int32Atom(shape=(), dflt=0), shape=(2,), pos=3)},
> "name": StringCol(itemsize=2, shape=(), dflt='', pos=3),
> "z2": UInt8Col(shape=(), dflt=1, pos=4)},
> "color": StringCol(itemsize=2, shape=(), dflt=' ', pos=2),
> "info": {
> "Name": StringCol(itemsize=2, shape=(), dflt='', pos=0),
> "Value": ComplexCol(itemsize=16, shape=(), dflt=(0.0+0.0j),
> pos=1)}, "y": Float64Col(shape=(2, 2), dflt=1.0, pos=4),
> "z": UInt8Col(shape=(), dflt=1, pos=5)}
>
>
> Test failed: <failed>
Yes, it seems like the representation for complex types will change in
1.0.5. I've opened a ticket in order to no forget fixing this in
PyTables tests:
http://www.pytables.org/trac/ticket/161
Cheers,
--
>0,0< Francesc Altet http://www.carabos.com/
V V Cárabos Coop. V. Enjoy Data
"-"
-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
Pytables-users mailing list
Pytables-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/pytables-users
