File Handling

As in all other good languages, Tcl also can open, read and write to files. And like all other good tutorials, this tutorial also teaches how to do it. First let us see how to open a file.

open

fileName is the name of the file. The access argument, if present, specifies the way in which the file should be accessed. It can have any of the following values:

#Open the file called "jokes.txt" for writing
open "jokes.txt" w

Now we know how to open a file. Not much good, is it? To make this function useful, we need to write to the file.

puts

Use the -nonewline option only if you don't need a new line at the end of the string you write to a file. The channelID stands for the ID of the output stream that must be written to(if you don't understand what the means, don't worry. You will get it latter.) string is the string you want to write to the file. Lets see the example.

#Open the file called "jokes.txt" for writing
set out [open "jokes.txt" w]
puts $out "Computers make very fast, very accurate mistakes."

Notice that I created a variable called 'out'. That will store the ID of the opened file. Always open a file in this way - otherwise they won't be of much use. Then we use that ID to write to the file using the command 'puts'. Now the file called "jokes.txt" has one line. Next we have to close the file.

close

This command closes the channel. Always do this - else you will cry later. Lets move on to the ever growing example...

#Open the file called "jokes.txt" for writing
set out [open "jokes.txt" w]
puts $out "Computers make very fast, very accurate mistakes."
close $out

After we close the file, we decide that we have to put more lines in the file. So now we open the file again, this time in the append mode. The example grows even longer...

#Open the file called "jokes.txt" for writing
set out [open "jokes.txt" w]
puts $out "Computers make very fast, very accurate mistakes."
close $out
set out [open "jokes.txt" a]
puts $out "Computers are not intelligent. They only think they are."
puts $out "My software never has bugs. It just develops random features."
puts $out {All computers wait at the same speed.
Best file compression around:  "DEL *.*" = 100% compression
DEFINITION: Computer - A device designed to speed and automate errors.
DEFINITION: Upgrade - Take old bugs out, put new ones in.}
close $out

After we have done that, we need to read the jokes and put them back on the screen. So we learn a new command. Ladies and Gentlemen let me introduce you to...

gets

gets will copy one line from the channel(or file) and put it in varName. If varName is not specified, the copied line will be the result of the function. We go back to our example and get the first line of the file. For that we open the file again, this time in read mode.

#Open the file called "jokes.txt" for writing
set out [open "jokes.txt" w]
puts $out "Computers make very fast, very accurate mistakes."
close $out
#Now append more jokes at the end of the file
set out [open "jokes.txt" a]
puts $out "Computers are not intelligent. They only think they are."
puts $out "My software never has bugs. It just develops random features."
puts $out {All computers wait at the same speed.
Best file compression around:  "DEL *.*" = 100% compression
DEFINITION: Computer - A device designed to speed and automate errors.
DEFINITION: Upgrade - Take old bugs out, put new ones in.}
close $out
#Opening file in read mode
set in [open "jokes.txt" r]
gets $in line
label .line -text "First Line : $line"
pack .line

This command can be used to read the whole file line by line. See the example below to see how. Don't worry about .txt insert end "$line\n---\n" - I will explain it later.

text .txt
set in [open "Jokes.txt" r]
while {[gets $in line] != -1} {
    #Do whatever you want with the $line variable
    .txt insert end "$line\n---\n"
}
close $in
pack .txt -expand 1 -fill both

This command is for reading the file line by line. Now we decide that we need to see the entire file. For that we need the read command. But remember that if we have read one line before, the read command will read only from the second line. For resetting the channel to the first position, we need the seek command.

seek

The offset and origin arguments specify the position at which the next read or write will occur for channelId. Offset must be an integer (which may be negative) and origin must be one of the following - start, current or end.

Now we have to read the file contents. So we move on to...

read

read will take numChars characters from the channel and return it. If numChars is not specified, the whole file is read and its content will be returned. You can read a file and then get all the lines into a list with every line as an item in the list. This can be done by...

set in [open "file.txt" r]
set contents [read $in]
close $in
set lines [split $contents "\n"]

Now every item in the list called $lines is a line of the file called "file.txt". On to the example - for the last time.

#Open the file called "jokes.txt" for writing
set out [open "jokes.txt" w]
puts $out "Computers make very fast, very accurate mistakes."
close $out
#Now append more jokes at the end of the file
set out [open "jokes.txt" a]
puts $out "Computers are not intelligent. They only think they are."
puts $out "My software never has bugs. It just develops random features."
puts $out {All computers wait at the same speed.
Best file compression around:  "DEL *.*" = 100% compression
DEFINITION: Computer - A device designed to speed and automate errors.
DEFINITION: Upgrade - Take old bugs out, put new ones in.}
close $out
#Opening file in read mode
set in [open "jokes.txt" r]
gets $in line
label .line -text "First Line : $line"
pack .line
seek $in 0 start 
set contents [read $in]
close $in
label .full -text "Full file Contents... \n$contents"
pack .full

There. That's about it with file handling. Now lets move on to juicier topics.