Pandas enh: api cleanup... tostring, to_string, to_html, etc.

There is a lot of semi-duplication / differences in the Series() and DataFrame() methods.

Maybe it's worth taking a look through them and regularising.

Fwiw, I prefer tostring vs. to_string

Also, would be nice to have tohtml on things, though expanding it is beyond the scope of this.

Comment From: wesm

It might be a bit too late to remove underscores from a host of methods. I find the underscores easier on the eyes, though. For example:

frame.to_panel()

looks much better to me than

frame.topanel()

but that may be largely a function of having used that convention for a long time.

I agree that the API should be made uniform and consistent across the data structures. Feel free to hack away-- I'm not going to be able to invest much in this relative to the feature queue for a while

Comment From: lodagro

As reference, with a little bit of pandas code i created an overview of available methods (0.8.0.dev-de6318e). Unfortunately github does not render colors, True/False are in Green/Red divs. True : method available False : method not available

	series	dataframe	panel
T	True	True	False
abs	True	True	True
add	True	True	True
add_prefix	False	True	True
add_suffix	False	True	True
align	True	True	False
all	True	False	False
any	True	False	False
append	True	True	False
apply	True	True	True
applymap	False	True	False
argmax	True	False	False
argmin	True	False	False
argsort	True	False	False
as_matrix	False	True	False
asfreq	True	True	False
asof	True	False	False
astype	True	True	True
autocorr	True	False	False
axes	False	True	True
base	True	False	False
boxplot	False	True	False
byteswap	True	False	False
choose	True	False	False
clip	True	True	False
clip_lower	True	True	False
clip_upper	True	True	False
columns	False	True	False
combine	True	True	False
combineAdd	False	True	False
combineMult	False	True	False
combine_first	True	True	False
compound	False	False	True
compress	True	False	False
conform	False	False	True
conj	True	False	False
conjugate	True	False	False
consolidate	False	True	True
convert_objects	False	True	False
copy	True	True	True
corr	True	True	False
corrwith	False	True	False
count	True	True	True
cov	True	True	False
ctypes	True	False	False
cummax	True	True	True
cummin	True	True	True
cumprod	True	True	True
cumsum	True	True	True
data	True	False	False
delevel	False	True	False
describe	True	True	False
diagonal	True	False	False
diff	True	True	False
div	True	True	True
divide	False	False	True
dot	True	True	False
drop	True	True	True
drop_duplicates	False	True	False
dropna	True	True	False
dtype	True	False	False
dtypes	False	True	False
dump	True	False	False
dumps	True	False	False
duplicated	False	True	False
fill	True	False	False
fillna	True	True	True
filter	False	True	True
first_valid_index	True	True	False
flags	True	False	False
flat	True	False	False
flatten	True	False	False
fromDict	False	False	True
from_csv	True	True	False
from_dict	False	True	True
from_items	False	True	False
from_records	False	True	False
get	True	True	True
get_dtype_counts	False	True	False
get_value	True	True	True
getfield	True	False	False
groupby	True	True	True
head	True	True	False
hist	True	True	False
icol	False	True	False
idxmax	True	True	False
idxmin	True	True	False
iget	True	False	False
iget_value	True	True	False
imag	True	False	False
index	True	True	False
info	False	True	False
insert	False	True	False
interpolate	True	False	False
irow	False	True	False
isin	True	False	False
isnull	True	False	False
item	True	False	False
items	False	False	True
itemset	True	False	False
itemsize	True	False	False
iteritems	True	True	True
iterkv	True	True	True
iterrows	False	True	False
ix	True	True	True
join	False	True	True
keys	True	False	True
last_valid_index	True	True	False
load	True	True	True
lookup	False	True	False
mad	True	True	False
major_axis	False	False	True
major_xs	False	False	True
map	True	False	False
max	True	True	True
mean	True	True	True
median	True	True	True
merge	False	True	False
min	True	True	True
minor_axis	False	False	True
minor_xs	False	False	True
mul	True	True	True
multiply	False	False	True
name	True	False	False
nbytes	True	False	False
ndim	True	True	True
newbyteorder	True	False	False
nonzero	True	False	False
notnull	True	False	False
nunique	True	False	False
order	True	False	False
pivot	False	True	False
pivot_table	False	True	False
plot	True	True	False
pop	False	True	True
prod	True	True	True
product	False	True	False
ptp	True	False	False
put	True	False	False
quantile	True	True	False
radd	False	True	False
rank	True	True	False
ravel	True	False	False
rdiv	False	True	False
real	True	False	False
reindex	True	True	True
reindex_axis	False	True	False
reindex_like	True	True	True
rename	True	True	False
rename_axis	False	True	True
reorder_levels	True	True	False
repeat	True	False	False
reset_index	True	True	False
reshape	True	False	False
resize	True	False	False
rmul	False	True	False
round	True	False	False
rsub	False	True	False
s	False	True	False
save	True	True	True
searchsorted	True	False	False
select	True	True	True
set_index	False	True	False
set_value	True	True	True
setasflat	True	False	False
setfield	True	False	False
setflags	True	False	False
shape	True	True	True
shift	True	True	True
size	True	False	False
skew	True	True	True
sort	True	True	False
sort_index	True	True	True
sortlevel	True	True	False
squeeze	True	False	False
stack	False	True	False
std	True	True	True
strides	True	False	False
sub	True	True	True
subtract	False	False	True
sum	True	True	True
swapaxes	True	False	True
swaplevel	True	True	True
tail	True	True	False
take	True	True	True
toLong	False	False	True
to_csv	True	True	False
to_dict	True	True	False
to_excel	False	True	True
to_frame	False	False	True
to_html	False	True	False
to_long	False	False	True
to_panel	False	True	False
to_records	False	True	False
to_sparse	True	True	True
to_string	True	True	False
to_wide	False	True	False
tofile	True	False	False
tolist	True	False	False
tostring	True	False	False
trace	True	False	False
transpose	True	True	False
truncate	True	True	True
unique	True	False	False
unstack	True	True	False
valid	True	False	False
value_counts	True	False	False
values	True	True	True
var	True	True	True
view	True	False	False
weekday	True	False	False
xs	False	True	True

Comment From: adamklein

Very useful and interesting!

Comment From: jreback

@lodagro can you update this? would be interesting now that objects are combined

Comment From: lodagro

Update? Do you mean regenerate the above table for master?

Comment From: jreback

yep...that's what I meant

Comment From: cpcloud

@lodagro can u also put the code u used to create it in a gist? we could make it into a script ...

Comment From: lodagro

Could not find back the code i used to generate the above table, i probably did it interactively using IPython.

Its not much really, on request i made a gist. The output is a html file. Note that you can use a predicate when doing inspect.getmembers to do some filtering.

Comment From: naught101

Dunno whether it's worth opening a new issue, but df.as_matrix should be renamed to df.as_array - it returns a numpy.ndarray, not a numpy.matrixlib.defmatrix.matrix. And in the case of Series, Panels and Panel4Ds it returns 1D, 3D and 4D arrays respectively, which clearly aren't matrices.

I would suggest that it might be worth adopting a convention like df.as_* for conversion to other data types (df.as_matrix, df.as_dict, df.as_list, etc), and df.to_* for export to files or other streams (df.to_clipboard, df.to_csv, df.to_html, etc).

Comment From: naught101

Also, I agree that underscores are better - what the hell is a tost ring? Not to mention "asof" (wtf?), and "top anel" :P

camelCase would also be ok, it's much clearer than no word definition, and it's shorter than underscores.

Re: "a bit too late". Wouldn't it be easy enough to just make aliases for each method, and raise warnings when they're called, noting that they're deprecated? Those aliases could be removed later (and the warnings could state which version they will be removed in).

Comment From: wesm

Closing -- please open issues addressing specific existing inconsistencies.