The following is using expl3 for the data processing. It works in the following way:
First read in the first line of the file and parse the header. Find the column which is called summits, so that in further processing we know which column is the interesting one.
Next read the file line by line and extract the $n$-th column (the one named summits). Split that column on commas (to get all the summits in that line), and then extract the name and height of each summit via a regular expression. Put the found values into a single seq variable.
Afterwards we remove duplicates from this seq, and sort for the height of the mountains (in descending order).
Then output the table. Each seq element in the now sorted table is given to \readmountains_output_single:n which formats every line.
This is all done by the macro \readmountains, which takes an optional and a mandatory argument. The mandatory argument is the file name, the optional one is the column delimiter (defaulting to a Tab, which is ^^I if you want to input it directly from within TeX, but the optional argument wants the delimiter escaped, so you'd provide \^^I for a Tab, which is the default).
(there is a bit of error checking going on, so that
\documentclass[]{article}
\begin{filecontents}{\jobname.csv}
title;summits;description
test1;Matterhorn (4500m);description1
test2;Mount Everest (8848m);description2
test3;K2 (8611m), Kangchenjunga (8586m);description3
test4;Mount Everest (8848m);description4
\end{filecontents}
\ExplSyntaxOn
\ior_new:N \g_readmountains_file_ior
\seq_new:N \l_readmountains_mountains_seq
\seq_new:N \l_readmountains_extract_seq
\int_new:N \g_readmountains_output_int
\int_new:N \l_readmountains_col_int
\int_new:N \l_readmountains_catcode_int
\int_new:N \l_readmountains_line_int
\seq_new:N \l_readmountains_csv_seq
\scan_new:N \s_readmountains_end
\msg_new:nnn { readmountains } { no-summits }
{ Column ~ summits' ~ not ~ found. ~ Aborting. } \msg_new:nnn { readmountains } { short-line } { Short ~ line ~ encountered ~ in ~ line ~ \int_use:N \l_readmountains_line_int. ~ Aborting. } \msg_new:nnn { readmountains } { malformatted-summit } { Malformatted ~ summit ~ encountered ~ in ~ line ~ \int_use:N \l_readmountains_line_int. ~ Aborting. } \NewDocumentCommand \readmountains { O{\^^I} m } { \readmountains:Nn #1 {#2} } \cs_new_protected:Npn \readmountains:Nn #1#2 { \readmountains_read_file:NnN #1 {#2} \l_readmountains_mountains_seq \seq_remove_duplicates:N \l_readmountains_mountains_seq \readmountains_sort:N \l_readmountains_mountains_seq \readmountains_output:N \l_readmountains_mountains_seq \use_none:n \s_readmountains_end } \cs_new_protected:Npn \readmountains_read_file:NnN #1#2#3 { \int_set:Nn \l_readmountains_catcode_int { \char_value_catcode:n {#1 } }
\char_set_catcode_other:N #1
\ior_open:Nn \g_readmountains_file_ior {#2}
\seq_clear:N #3
\ior_get:NN \g_readmountains_file_ior \l_tmpa_tl
\exp_args:NNx \seq_set_split:NnV \l_readmountains_csv_seq
{ \char_generate:nn { #1 } { 12 } } \l_tmpa_tl \int_zero:N \l_readmountains_col_int \seq_map_inline:Nn \l_readmountains_csv_seq { \int_incr:N \l_readmountains_col_int \str_if_eq:nnT { summits } {##1} { \seq_map_break: } } \str_if_eq:eeF { \seq_item:Nn \l_readmountains_csv_seq \l_readmountains_col_int } { summits } { \msg_error:nn { readmountains } { no-summits } \readmountains_use_none_delimit_by_s_end:w } \int_set:Nn \l_readmountains_line_int \c_one_int \ior_map_inline:Nn \g_readmountains_file_ior { \int_incr:N \l_readmountains_line_int \exp_args:NNx \seq_set_split:Nnn \l_readmountains_csv_seq { \char_generate:nn {#1 } { 12 } }
{##1}
\int_compare:nNnT
\l_readmountains_col_int > { \seq_count:N \l_readmountains_csv_seq }
{
\msg_error:nn { readmountains } { short-line }
\ior_map_break:n { \readmountains_use_none_delimit_by_s_end:w }
}
\exp_args:Nnnx
\seq_set_split:Nnn \l_readmountains_csv_seq {,}
{ \seq_item:Nn \l_readmountains_csv_seq \l_readmountains_col_int }
\seq_map_inline:Nn \l_readmountains_csv_seq
{
\tl_if_empty:nF { ####1 }
{
\regex_extract_once:nnNF { (.*)\s((\d+)m) } { ####1 }
\l_readmountains_extract_seq
{
\msg_error:nn { readmountains } { malformatted-summit }
\readmountains_use_none_delimit_by_s_end:w
}
\seq_push:Nx #3
{
{ \seq_item:Nn \l_readmountains_extract_seq {2} }
{ \seq_item:Nn \l_readmountains_extract_seq {3} }
}
\use_none:n \s_readmountains_end
}
}
}
\ior_close:N \g_readmountains_file_ior
\char_set_catcode:nn { `#1 } \l_readmountains_catcode_int
}
\cs_new_protected:Npn \readmountains_sort:N #1
{
\seq_sort:Nn #1
{
\int_compare:nNnTF { \use_ii:nn ##1 } < { \use_ii:nn ##2 }
\sort_return_swapped:
\sort_return_same:
}
}
\cs_new_protected:Npn \readmountains_output:N #1
{
\int_gzero:N \g_readmountains_output_int
\begin { tabular } { r l r }
\seq_map_function:NN #1 \readmountains_output_single:n
\end { tabular }
}
\cs_new_protected:Npn \readmountains_output_single:n #1
{
\int_gincr:N \g_readmountains_output_int
\int_use:N \g_readmountains_output_int .
& \use_i:nn #1
& \use_ii:nn #1
\
}
\cs_new:Npn \readmountains_use_none_delimit_by_s_end:w #1 \s_readmountains_end
{}
\ExplSyntaxOff
\begin{document}
% since the file was written with ; as the delimiter use that
\readmountains[;]{\jobname.csv}
\end{document}
;or,as the column delimiter. – Skillmon Jan 16 '21 at 19:25titleanddescriptionimportant or could they be dropped? – Skillmon Jan 16 '21 at 19:32