#2874 Set table deleting empty lines + mess with separator

[Anonymous]

Hey,

when using

set table
    plot ... w table
unset table

both white lines used to separate data blocks are lost.
Omitting "w table" reduces >=2 empty lines to one in the ou. In this case, I can work around by using

plot for [i=0:N] ... i i

but this doesn't work for active "w table" option.
I guess this is not intended behaviour?

Also, I didn't find a good way to combine plotting data from a file with tab as explicit separator with data created using a table using smooth options: the smooth option can't be combined with "w table", whereas "set table separator tab" only works with the "w table" option.
Is there a reason why "set table" doesn't use the default/defined separator?

Tested on gnuplot v6.0 p4 on windows.

Many thanks

[Ethan Merritt]

I agree it is reasonable to preserve blank lines when processing by plot ... with table.
That much is easily added, and works as expected for, e.g., polygon data where the individual polygons are separated by a single blank line. Preserving double blank lines also works so long as the entire input file is processed in one go. However if the input is filtered by data set (the index keyword) then the double blanks may no longer be there. This might also be fixable but I do not immediately see how. In such cases it would probably be better to save the original data set number as an extra column and then refer to it later by referencing that column. Example
# select data sets 0 2 4 6 8
set table $TEMP
plot for [i=0:9:2] 'infile.dat' index i using 1:2:(column(-2)) with table
unset table
# $TEMP now contains 5 data sets but separated by only a single blank
# To plot the original 4th data set (original index 3) one could now do
plot $TEMP using 1:2 if (column(3)==3)

Fixed, will be in 6.0.5

I do not understand your question regarding interaction of tab separators and smoothing. Could you maybe open a new tracker item and provide a test case?