gnuplot-info

[Gnuplot-info] set encoding and data file input

[Philipp K. J. <ja...@ie...>]

I understand that choosing the correct encoding 
is important when creating text labels (and so
on) with characters outside the ASCII range.

But how does gnuplot treat non-ASCII characters
on input, that is, when reading a file in
	plot "file" u 3:4

I guess it's somewhat of a non-issue, as long as
all the data in the file is numeric - that can 
be represented with strict ASCII. 

But what if the file contains string columns -
do they get interpreted differently, based on
set encoding? And can I generally rely on 
gnuplot to handle multi-byte characters safely
(that is, recognize them as multibyte and store
them accordingly)?

Best,

		Ph.

Re: [Gnuplot-info] set encoding and data file input

[sfeam <sf...@us...>]

On Tuesday, 28 October 2014 10:45:48 PM Philipp K. Janert wrote:
> 
> I understand that choosing the correct encoding 
> is important when creating text labels (and so
> on) with characters outside the ASCII range.
> 
> But how does gnuplot treat non-ASCII characters
> on input, that is, when reading a file in
> 	plot "file" u 3:4
> 
> I guess it's somewhat of a non-issue, as long as
> all the data in the file is numeric - that can 
> be represented with strict ASCII. 
> 
> But what if the file contains string columns -
> do they get interpreted differently, based on
> set encoding? 

Yes and no.
The individual "words" will get read in and stored no matter
what the encoding.  But the program needs to know the encoding
in order to operate on them properly.  For example:

set encoding iso_8859_1    # This is wrong
test = "αβγδεζηθ"
print test			# This works because the whole thing is spit out
print test[2:2]		# This fails because the encoding is wrong
print test[3:4]		# As does this

set encoding utf8			# this fixes it
print test[2:2]
print test [3:4]


> And can I generally rely on 
> gnuplot to handle multi-byte characters safely
> (that is, recognize them as multibyte and store
> them accordingly)?

UTF8 - yes
shift-JIS - I think so but I don't use it myself
other - untested

> Best,
> 
> 		Ph.
> 
> 
> 
> ------------------------------------------------------------------------------
> _______________________________________________
> gnuplot-info mailing list
> gnu...@li...
> Membership management via: https://lists.sourceforge.net/lists/listinfo/gnuplot-info

Re: [Gnuplot-info] set encoding and data file input

[Philipp K. J. <ja...@ie...>]

[snip]

> > 
> > But what if the file contains string columns -
> > do they get interpreted differently, based on
> > set encoding? 
> 
> Yes and no.
> The individual "words" will get read in and stored no matter
> what the encoding.  But the program needs to know the encoding
> in order to operate on them properly.  For example:
> 
> set encoding iso_8859_1    # This is wrong
> test = "αβγδεζηθ"
> print test			# This works because the whole
> thing is spit out print test[2:2]		# This fails because
> the encoding is wrong print test[3:4]		# As does this
> 
> set encoding utf8			# this fixes it
> print test[2:2]
> print test [3:4]
> 
> 
> > And can I generally rely on 
> > gnuplot to handle multi-byte characters safely
> > (that is, recognize them as multibyte and store
> > them accordingly)?
> 
> UTF8 - yes
> shift-JIS - I think so but I don't use it myself
> other - untested
> 

Many thanks, very helpful!