I use the following command to sort the content of a string
set local_object [lsort -dictionary $list_object]
this comand will replace new lines by spaces
how to avoid that ?
lsort assumes that its argument is a Tcl list. Any whitespace including newlines can separate elements of that list, but will not be preserved in the output. If you want to format the output list with one element per line you could do:
set local_object [join [lsort -dictionary $list_object] "\n"]
It all depends on how your list is built. Any string can be interpreted as list. All the white spaces are considered as a delimiter if you're treating string as list.
set str "d b a\n c"
set lst [lsort -dictionary [split $str " "]]
foreach word $lst {
puts $word
}
a
b
c
d
Split has preserved newline and used single space as a delimiter.
Related
I have this string:
1, RotD50, 88, 0.1582, 1.2264, -, 7.4, 23.6, 0.2, "San Fernando", 1971, "Santa Felita Dam (Outlet)", 6.61, Reverse, 24.69, 24.87, 389.0, 0.125, 1.2939, RSN88_SFERN_FSD172.AT2, RSN88_SFERN_FSD262.AT2 , RSN88_SFERN_FSD-UP.AT2
I want to find the indices of RSN88_SFERN_FSD172.AT2 and RSN88_SFERN_FSD262.AT2
I have tried a few scripts (like the following) but want to see if someone can help me with a rigorous script?
set currentdirc [pwd]
set fp [open _SearchResults.csv]
set count 1
foreach line [split [read $fp] \n] {
foreach word [split $line] {
set word [string trim $word ","]
set index [lsearch -exact $word "Horizontal-1 Acc.Filename"]
puts "$index"
}
}
You are going to need this:
package require csv
As before, break the data into lines and iterate over those lines. Trim the data first to avoid empty lines before or after.
foreach line [split [string trim [read $fp]] \n] {
Instead of trying to split the csv data using the split command, use the dedicated command ::csv::split from the csv package in Tcllib. You probably have it in your Tcl installation already.
set words [::csv::split $line]
When your line is split, there is unwanted whitespace around many data fields. Let's trim it off.
set words [lmap word $words {string trim $word}]
Finally, you can search for data in the list of words. Searching in each word as you did is pointless.
set index [lsearch $words RSN88_SFERN_FSD262.AT2]
Putting it together:
foreach line [split [string trim [read $fp]] \n] {
set words [::csv::split $line]
set words [lmap word $words {string trim $word}]
set index [lsearch $words RSN88_SFERN_FSD262.AT2]
puts $index
}
Documentation:
csv (package),
foreach,
lmap (for Tcl 8.5),
lmap,
lsearch,
package,
puts,
read,
set,
split,
string
I would use the csv package to do that task, since you are dealing with a csv file. Splitting blindly will split 1, RotD50, 88, 0.1582, 1.2264, -, 7.4, 23.6, 0.2, "San Fernando" things into, for example (each element on their own line):
1,
RotD50,
88,
0.1582,
1.2264,
-,
7.4,
23.6,
0.2,
"San
Fernando"
So my suggestion is:
set currentdirc [pwd]
set fp [open [file join $currentdirc _SearchResults.csv] r]
package require csv
foreach line [split [read $fp] \n] {
set words [::csv::split $line]
set index [lsearch -exact $words "Horizontal-1 Acc.Filename"]
puts $index
}
Also the list of words is the whole line. So if you want to loop through the words, you would do if {$word eq "Horizontal-1 Acc.Filename"} instead and you would have to use count (that I removed in my suggestion) to keep track of the index.
If for some reason you cannot use the csv package, you can try using this instead of the line containing ::csv::split:
set all [regexp -all -inline -- {\"[^\"]+\"|[^,]+} $line]
set words [lmap w $all {set w [string trim $w {\" }]}]
(I'm using \" for the quotes only for the sake of proper syntax highlighting, you can safely use " alone)
I want to do two things:
1) count the number of times a given word appears in a text file
2) print out the context of that word
This is the code I am currently using:
my $word_delimiter = qr{
[^[:alnum:][:space:]]*
(?: [[:space:]]+ | -- | , | \. | \t | ^ )
[^[:alnum:]]*
}x;
my $word = "hello";
my $count = 0;
#
# here, a file's contents are loaded into $lines, code not shown
#
$lines =~ s/\R/ /g; # replace all line breaks with blanks (cannot just erase them, because this might connect words that should not be connected)
$lines =~ s/\s+/ /g; # replace all multiple whitespaces (incl. blanks, tabs, newlines) with single blanks
$lines = " ".$lines." "; # add a blank at beginning and end to ensure that first and last word can be found by regex pattern below
while ($lines =~ m/$word_delimiter$word$word_delimiter/g ) {
++$count;
# here, I would like to print the word with some context around it (i.e. a few words before and after it)
}
Three problems:
1) Is my $word_delimiter pattern catching all reasonable characters I can expect to separate words? Of course, I would not want to separate hyphenated words, etc. [Note: I am using UTF-8 throughout but only English and German text; and I understand what reasonably separates a word might be a matter of judgment]
2) When the file to be analzed contains text like "goodbye hello hello goodbye", the counter is incremented only once, because the regex only matches the first occurence of " hello ". After all, the second time it could find "hello", it is not preceeded by another whitespace. Any ideas on how to catch the second occurence, too? Should I maybe somehow reset pos()?
3) How to (reasonably efficiently) print out a few words before and after any matched word?
Thanks!
1. Is my $word_delimiter pattern catching all reasonable characters I can expect to separate words?
Word characters are denoted by the character class \w. It also matches digits and characters from non-roman scripts.
\W represents the negated sense (non-word characters).
\b represents a word boundary and has zero-length.
Using these already available character classes should suffice.
2. Any ideas on how to catch the second occurence, too?
Use zero-length word boundaries.
while ( $lines =~ /\b$word\b/g ) {
++$count;
}
I have a string as follows
set temp "
temp : value
temp1 : value1
tempvalue
abc = 23:445:726
abcdef = 456::985
abcdef = 123:45:7
abcdef = 098:45:56:8
"
In this I want an output in which the values after "=" should be set to one variable. Output should be
"456::985 123:45:7 098:45:56:8".
I used
set result [regexp "abcdef\\s*=\\s*(\\S*)" $temp match v1]
but not able to get all
I have got the answer using regexp with -inline -all and -line, to store the result in list and then traverse through it to get the values. I need a one liner
set result [regexp -inline -all -line "abcdef\\s*=\\s*(\\S*)" $temp]
Output is
{abcdef = 456::985} 456::985 {abcdef = 123:45:7} 123:45:7 {abcdef = 098:45:56:8} 098:45:56:8
Then traverse through this to set them all in one string. But i want to know if there is any easy way to do this.
Thanks in advance.
Given this example you don't need regexp. Split the lines into pieces and create a new list.
set r {}
foreach line [split $temp \n] {
if {[string trim $line] eq ""} continue; # skip blank lines
lappend r [string trim [lindex [split $line =] end]]
}
puts $r
That will give one list with just the bits after the equals sign. If you treat it as a string, then it works as a string with each element of the list separated by a space.
Here is another approach: think of each line as 3 tokens: key, equal sign, and value:
set result {}
foreach {key eq_sign value} $temp { lappend result $value }
This approach is simple to understand, but it will not work if the value contains spaces.
Since you're using line-oriented matching, take advantage of the line anchor:
% regexp -inline -all -line {\S+$} $temp
456::985 123:45:7 098:45:56:8
So, to save the values as a string:
set values [join [regexp -inline -all -line {\S+$} $temp]]
If there may not be whitespace after the equal, use the pattern {[^=\s]+$}
My whole answer assumes there's only going to be one equal sign per line.
Responding to the updated sample input:
foreach {m word} [regexp -inline -all -line {=\s*(\S+)$} $temp] {
lappend words $word
}
puts [join $words]
23:445:726 456::985 123:45:7 098:45:56:8
I was originally puzzled by this: I was just working on a character splitting function in Perl, when I noticed this:
DB<56> map(print("-", $_, "\n"), split( //, "test") );
-t
-e
-s
-t
DB<57> map(print("-", $_, "\n"), split( /./, "test") );
DB<58> map(print("-", $_, "\n"), split( /(.)/, "test") );
-
-t
-
-e
-
-s
-
-t
I already knew that if the empty regex // is used, the string is split into individual characters; but I wasn't clear on where did those empty strings in the /(.)/ regex come from - but just a few sentences later, the page states "If the regex has groupings, then the list produced contains the matched substrings from the groupings as well ... Since the first character of $x matched the regex, split prepended an empty initial element to the list." So, it's expected behavior. (althgouh, I'm still not clear why ungrouped dot /./ doesn't do anything )
But, I was also working in Python, and encountered a similar problem (empty strings in result of split) - and there I found a filter(None, list) function, which in this invocation, simply removes empty strings from a list. What is used to achieve the same in Perl?
The first argument of split defines what separates the terms of the list you are parsing. In your last two snippets, you tell split that any character is a valid separator, so split returns what's between the characters of the input: Five empty strings.
>perl -E"say qq{<$_>} for split /./, 'test', -1;"
<>
<>
<>
<>
<>
(Trailing empty strings are filtered out by default.)
The solution is not to start filtering out the very thing you asked split to produce. Either fix your separator
my #chars = split /(?<=.)|(?=.)/s;
my #chars = split //;
or use a better tool
my #chars = /(.)/s;
my #chars = unpack '(a)*', $_;
I am trying to extract a substring from a string in Tcl. I wrote the code and able to do it, but I was wondering if there is any other efficient way to do it. So the exact problem is I have a string
name_ext_10a.string_10a.string.string.string
and I want to extract "name_ext", and then remove that "_" and replace it with "."; I finally want the output to be "name.ext". I wrote something like this:
set _File "[string replace $_File [string last "_" $_File] [string length $_File] "" ]"
set _File "[string replace $_File [string last "_" $_File] [string length $_File] "" ]"
set _File "[string replace $_File [string last "_" $_File] [string last "_" $_File] "." ]"
which gives me the exact output I want, but I was wondering if there is any other efficient way to do this in Tcl.
You could split that filename using underscore as a separator, and then join the first 2 elements with a dot:
% set f name_ext_10a.string_10a.string.string.string
name_ext_10a.string_10a.string.string.string
% set out [join [lrange [split $f _] 0 1] .]
name.ext
EDIT
So if "name" can have an arbitrary number of underscores:
set f "foo_bar_baz_ext_10a.string_10a.string.string.string"
set pieces [split $f _]
set name [join [lrange $pieces 0 end-3] _]
set out [join [list $name [lindex $pieces end-2]] .] ;#==> foo_bar_baz.ext
But this is getting complex. One regex should suffice -- I assume "string" can be any sequence of non-underscore chars.
set string {[^_]+}
set regex "^(.+)_($string)_10a.${string}_10a.$string.$string.$string\$"
regexp $regex $f -> name ext
set out "$name.$ext" ;#==> foo_bar_baz.ext
One way to do the extraction is with regsub:
regsub {^([^_]+)_([^_]+)_.*} $_File {\1.\2} _File
The regular expression contains ([^_]+) components, which match a sequence of non-underscore characters, plus an anchor and some underscores, and a trailing non-capturing .* which matches everything else (so we can discard it). The regsub replaces that (which is the whole string) with the concatenation of the two matched non-underscore sections with a . between, and writes it back to the _File variable where the string came from.
Note that I put the regular expression and replacement in braces. This is because they contain Tcl metacharacters (square brackets and backslashes) which I want Tcl to pass into regsub verbatim.