Read a file character by character/UTF8
![Task](http://static.miraheze.org/rosettacodewiki/thumb/b/ba/Rcode-button-task-crushed.png/64px-Rcode-button-task-crushed.png)
You are encouraged to solve this task according to the task description, using any language you may know.
- Task
Read a file one character at a time, as opposed to reading the entire file at once.
The solution may be implemented as a procedure, which returns the next character in the file on each consecutive call (returning EOF when the end of the file is reached).
The procedure should support the reading of files containing UTF8 encoded wide characters, returning whole characters for each consecutive read.
- Related task
AutoHotkey
<lang AutoHotkey>File := FileOpen("input.txt", "r") while !File.AtEOF
MsgBox, % File.Read(1)</lang>
C
<lang C>#include <stdio.h>
- include <wchar.h>
- include <stdlib.h>
- include <locale.h>
int main(void) {
/* If your native locale doesn't use UTF-8 encoding * you need to replace the empty string with a * locale like "en_US.utf8" */ char *locale = setlocale(LC_ALL, ""); FILE *in = fopen("input.txt", "r");
wint_t c; while ((c = fgetwc(in)) != WEOF) putwchar(c); fclose(in);
return EXIT_SUCCESS;
}</lang>
C#
<lang csharp>using System; using System.IO; using System.Text;
namespace RosettaFileByChar {
class Program { static char GetNextCharacter(StreamReader streamReader) => (char)streamReader.Read();
static void Main(string[] args) { Console.OutputEncoding = Encoding.UTF8; char c; using (FileStream fs = File.OpenRead("input.txt")) { using (StreamReader streamReader = new StreamReader(fs, Encoding.UTF8)) { while (!streamReader.EndOfStream) { c = GetNextCharacter(streamReader); Console.Write(c); } } } } }
}</lang>
Common Lisp
<lang lisp>;; CLISP puts the external formats into a separate package
- +clisp (import 'charset:utf-8 'keyword)
(with-open-file (s "input.txt" :external-format :utf-8)
(loop for c = (read-char s nil) while c do (format t "~a" c)))</lang>
Crystal
The encoding is UTF-8 by default, but it can be explicitly specified.
<lang ruby>File.open("input.txt") do |file|
file.each_char { |c| p c }
end</lang>
or
<lang ruby>File.open("input.txt") do |file|
while c = file.read_char p c end
end</lang>
Delphi
<lang Delphi> program Read_a_file_character_by_character_UTF8;
{$APPTYPE CONSOLE}
uses
System.SysUtils, System.Classes;
function GetNextCharacter(StreamReader: TStreamReader): char; begin
Result := chr(StreamReader.Read);
end;
const
FileName: TFileName = 'input.txt';
begin
if not FileExists(FileName) then raise Exception.Create('Error: File not exist.');
var F := TStreamReader.Create(FileName, TEncoding.UTF8);
while not F.EndOfStream do begin var c := GetNextCharacter(F); write(c); end; readln;
end.</lang>
Déjà Vu
<lang dejavu>#helper function that deals with non-ASCII code points local (read-utf8-char) file tmp:
!read-byte file if = :eof dup: drop raise :unicode-error resize-blob tmp ++ dup len tmp set-to tmp try: return !decode!utf-8 tmp catch unicode-error: if < 3 len tmp: raise :unicode-error (read-utf8-char) file tmp
- reader function
read-utf8-char file:
!read-byte file if = :eof dup: return local :tmp make-blob 1 set-to tmp 0 try: return !decode!utf-8 tmp catch unicode-error: (read-utf8-char) file tmp
- if the module is used as a script, read from the file "input.txt",
- showing each code point separately
if = (name) :(main):
local :file !open :read "input.txt"
while true: read-utf8-char file if = :eof dup: drop !close file return !.</lang>
Factor
<lang>USING: kernel io io.encodings.utf8 io.files strings ; IN: rosetta-code.read-one
"input.txt" utf8 [
[ read1 dup ] [ 1string write ] while drop
] with-file-reader</lang>
FunL
<lang funl>import io.{InputStreamReader, FileInputStream}
r = InputStreamReader( FileInputStream('input.txt'), 'UTF-8' )
while (ch = r.read()) != -1
print( chr(ch) )
r.close()</lang>
Go
<lang go>package main
import (
"bufio" "fmt" "io" "os"
)
func Runer(r io.RuneReader) func() (rune, error) {
return func() (r rune, err error) { r, _, err = r.ReadRune() return }
}
func main() {
runes := Runer(bufio.NewReader(os.Stdin)) for r, err := runes(); err != nil; r,err = runes() { fmt.Printf("%c", r) }
}</lang>
Haskell
<lang haskell>#!/usr/bin/env runhaskell
{- The procedure to read a UTF-8 character is just:
hGetChar :: Handle -> IO Char
assuming that the encoding for the handle has been set to utf8.
-}
import System.Environment (getArgs) import System.IO (
Handle, IOMode (..), hGetChar, hIsEOF, hSetEncoding, stdin, utf8, withFile )
import Control.Monad (forM_, unless) import Text.Printf (printf) import Data.Char (ord)
processCharacters :: Handle -> IO () processCharacters h = do
done <- hIsEOF h unless done $ do c <- hGetChar h putStrLn $ printf "U+%04X" (ord c) processCharacters h
processOneFile :: Handle -> IO () processOneFile h = do
hSetEncoding h utf8 processCharacters h
{- You can specify one or more files on the command line, or if no
files are specified, it reads from standard input.
-} main :: IO () main = do
args <- getArgs case args of [] -> processOneFile stdin xs -> forM_ xs $ \name -> do putStrLn name withFile name ReadMode processOneFile</lang>
- Output:
bash$ echo €50 | ./read-char-utf8.hs U+20AC U+0035 U+0030 U+000A
J
Reading a file a character at a time is antithetical not only to the architecture of J, but to the architecture and design of most computers and most file systems. Nevertheless, this can be a useful concept if you're building your own hardware. So let's model it...
First, we know that the first 8-bit value in a utf-8 sequence tells us the length of the sequence needed to represent that character. Specifically: we can convert that value to binary, and count the number of leading 1s to find the length of the character (except the length is always at least 1 character long).
<lang J>u8len=: 1 >. 0 i.~ (8#2)#:a.&i.</lang>
So now, we can use indexed file read to read a utf-8 character starting at a specific file index. What we do is read the first octet and then read as many additional characters as we need based on whatever we started with. If that's not possible, we will return EOF:
<lang J>indexedread1u8=:4 :0
try. octet0=. 1!:11 y;x,1 octet0,1!:11 y;(x+1),<:u8len octet0 catch. 'EOF' end.
)</lang>
The length of the result tells us what to add to the file index to find the next available file index for reading.
Of course, this is massively inefficient. So if someone ever asks you to do this, make sure you ask them "Why?" Because the answer to that question is going to be important (and might suggest a completely different implementation).
Note also that it would make more sense to return an empty string, instead of the string 'EOF', when we reach the end of the file. But that is out of scope for this task.
Java
<lang java>import java.io.FileReader; import java.io.IOException; import java.nio.charset.StandardCharsets;
public class Main {
public static void main(String[] args) throws IOException { var reader = new FileReader("input.txt", StandardCharsets.UTF_8); while (true) { int c = reader.read(); if (c == -1) break; System.out.print(Character.toChars(c)); } }
}</lang>
jq
jq being stream-oriented, it makes sense to define `readc` so that it emits a stream of the UTF-8 characters in the input: <lang jq>def readc:
inputs + "\n" | explode[] | [.] | implode;</lang>
Example: <lang sh>
echo '过活' | jq -Rn 'include "readc"; readc' "过" "活" "\n"</lang>
Julia
The built-in read(stream, Char)
function reads a single UTF8-encoded character from a given stream.
<lang julia>open("myfilename") do f
while !eof(f) c = read(f, Char) println(c) end
end</lang>
Kotlin
<lang scala>// version 1.1.2
import java.io.File
const val EOF = -1
fun main(args: Array<String>) {
val reader = File("input.txt").reader() // uses UTF-8 by default reader.use { while (true) { val c = reader.read() if (c == EOF) break print(c.toChar()) // echo to console } }
}</lang>
Lua
<lang Lua> -- Return whether the given string is a single ASCII character. function is_ascii (str)
return string.match(str, "[\0-\x7F]")
end
-- Return whether the given string is an initial byte in a multibyte sequence. function is_init (str)
return string.match(str, "[\xC2-\xF4]")
end
-- Return whether the given string is a continuation byte in a multibyte sequence. function is_cont (str)
return string.match(str, "[\x80-\xBF]")
end
-- Accept a filestream. -- Return the next UTF8 character in the file. function read_char (file)
local multibyte -- build a valid multibyte Unicode character
for c in file:lines(1) do if is_ascii(c) then if multibyte then -- We've finished reading a Unicode character; unread the next byte, -- and return the Unicode character. file:seek("cur", -1) return multibyte else return c end elseif is_init(c) then if multibyte then file:seek("cur", -1) return multibyte else multibyte = c end elseif is_cont(c) then multibyte = multibyte .. c else assert(false) end end
end
-- Test. function read_all ()
testfile = io.open("tmp.txt", "w") testfile:write("𝄞AöЖ€𝄞Ελληνικάy䮀成长汉\n") testfile:close() testfile = io.open("tmp.txt", "r")
while true do local c = read_char(testfile) if not c then return else io.write(" ", c) end end
end </lang>
- Output:
𝄞 A ö Ж € 𝄞 Ε λ λ η ν ι κ ά y ä ® € 成 长 汉
M2000 Interpreter
from revision 27, version 9.3, of M2000 Environment, Chinese 长 letter displayed in console (as displayed in editor)
<lang M2000 Interpreter> Module checkit {
\\ prepare a file \\ Save.Doc and Append.Doc to file, Load.Doc and Merge.Doc from file document a$ a$={First Line Second line Third Line Ελληνικά Greek Letters y䮀 成长汉 } Save.Doc a$, "checkthis.txt", 2 ' 2 for UTF-8 b$="*" final$="" buffer Clear bytes as byte*16 Buffer One as byte Buffer Two as byte*2 Buffer Three as byte*3 Locale 1033 open "checkthis.txt" for input as #f seek#f, 4 ' skip BOM While b$<>"" { GetOneUtf8Char(&b$) final$+=b$ } close #f Report final$ Sub GetOneUtf8Char(&ch$) ch$="" if Eof(#f) then Exit Sub Get #f, One Return Bytes, 0:=Eval(one, 0) local mrk=Eval(one, 0) Try ok { If Binary.And(mrk, 0xE0)=0xC0 then { Get #f,one Return Bytes, 1:=Eval$(one, 0,1) ch$=Eval$(Bytes, 0, 2) } Else.if Binary.And(mrk, 0xF0)=0xE0 then { Get #f,two Return Bytes, 1:=Eval$(two,0,2) ch$=Eval$(Bytes, 0, 3) } Else.if Binary.And(mrk, 0xF8)=0xF0 then { Get #f,three Return Bytes, 1:=Eval$(three, 0, 3) ch$=Eval$(Bytes, 0, 4) } Else ch$=Eval$(Bytes, 0, 1) } if Error or not ok then ch$="" : exit sub ch$=left$(string$(ch$ as Utf8dec),1) End Sub
} checkit </lang>
using document as final$
<lang M2000 Interpreter> Module checkit {
\\ prepare a file \\ Save.Doc and Append.Doc to file, Load.Doc and Merge.Doc from file document a$ a$={First Line Second line Third Line Ελληνικά Greek Letters y䮀 成长汉 } Save.Doc a$, "checkthis.txt", 2 ' 2 for UTF-8 b$="*" document final$ buffer Clear bytes as byte*16 Buffer One as byte Buffer Two as byte*2 Buffer Three as byte*3 Locale 1033 open "checkthis.txt" for input as #f seek#f, 4 ' skip BOM oldb$="" While b$<>"" { GetOneUtf8Char(&b$) \\ if final$ is document then 10 and 13 if comes alone are new line \\ so we need to throw 10 after the 13, so we have to use oldb$ if b$=chr$(10) then if oldb$=chr$(13) then oldb$="": continue oldb$=b$ final$=b$ ' we use = for append to document } close #f Report final$ Sub GetOneUtf8Char(&ch$) ch$="" if Eof(#f) then Exit Sub Get #f, One Return Bytes, 0:=Eval(one, 0) local mrk=Eval(one, 0) Try ok { If Binary.And(mrk, 0xE0)=0xC0 then { Get #f,one Return Bytes, 1:=Eval$(one, 0,1) ch$=Eval$(Bytes, 0, 2) } Else.if Binary.And(mrk, 0xF0)=0xE0 then { Get #f,two Return Bytes, 1:=Eval$(two,0,2) ch$=Eval$(Bytes, 0, 3) } Else.if Binary.And(mrk, 0xF8)=0xF0 then { Get #f,three Return Bytes, 1:=Eval$(three, 0, 3) ch$=Eval$(Bytes, 0, 4) } Else ch$=Eval$(Bytes, 0, 1) } if Error or not ok then ch$="" : exit sub ch$=left$(string$(ch$ as Utf8dec),1) End Sub
} checkit
</lang>
Mathematica/Wolfram Language
<lang Mathematica>str = OpenRead["file.txt"]; ToString[Read[str, "Character"], CharacterEncoding -> "UTF-8"]</lang>
NetRexx
Java and by extension NetRexx provides I/O functions that read UTF-8 encoded character data directly from an attached input stream. The Reader.read() method reads a single character as an integer value in the range 0 – 65535 [0x00 – 0xffff], reading from a file encoded in UTF-8 will read each codepoint into an int. In the sample below the readCharacters method reads the file character by character into a String and returns the result to the caller. The rest of this sample examines the result and formats the details.
- The file data/utf8-001.txt it a UTF-8 encoded text file containing the following: y䮀𝄞𝄢12.
<lang NetRexx>/* NetRexx */ options replace format comments java crossref symbols nobinary numeric digits 20
runSample(arg) return
-- ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ method readCharacters(fName) public static binary returns String
slurped = String() slrp = StringBuilder() fr = Reader null fFile = File(fName) EOF = int -1 -- End Of File indicator do fr = BufferedReader(FileReader(fFile)) ic = int cc = char -- read the contents of the file one character at a time loop label rdr forever -- Reader.read reads a single character as an integer value in the range 0 - 65535 [0x00 - 0xffff] -- or -1 on end of stream i.e. End Of File ic = fr.read() if ic == EOF then leave rdr cc = Rexx(ic).d2c slrp.append(cc) end rdr -- load the results of the read into a variable slurped = slrp.toString() catch fex = FileNotFoundException fex.printStackTrace() catch iex = IOException iex.printStackTrace() finally if fr \= null then do fr.close() catch iex = IOException iex.printStackTrace() end end return slurped
-- ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ method encodingDetails(str = String) public static
stlen = str.length() cplen = Character.codePointCount(str, 0, stlen) say 'Unicode: length="'stlen'" code_point_count="'cplen'" string="'str'"' loop ix = 0 to stlen - 1 cp = Rexx(Character.codePointAt(str, ix)) cc = Rexx(Character.charCount(cp)) say ' 'formatCodePoint(ix, cc, cp) if cc > 1 then do surrogates = [Rexx(Character.highSurrogate(cp)).c2d(), Rexx(Character.lowSurrogate(cp)).c2d()] loop sx = 0 to cc - 1 ix = ix + sx cp = surrogates[sx] say ' 'formatCodePoint(ix, 1, cp) end sx end end ix say return
-- ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ -- @see http://docs.oracle.com/javase/6/docs/technotes/guides/intl/encoding.doc.html -- @since Java 1.7 method formatCodePoint(ix, cc, cp) private static
scp = Rexx(Character.toChars(cp)) icp = cp.d2x(8).x2d(9) -- signed to unsigned conversion ocp = Rexx(Integer.toOctalString(icp)) x_utf16 = x_utf8 = do b_utf16 = String(scp).getBytes('UTF-16BE') b_utf8 = String(scp).getBytes('UTF-8') loop bv = 0 to b_utf16.length - 1 by 2 x_utf16 = x_utf16 Rexx(b_utf16[bv]).d2x(2) || Rexx(b_utf16[bv + 1]).d2x(2) end bv loop bv = 0 to b_utf8.length - 1 x_utf8 = x_utf8 Rexx(b_utf8[bv]).d2x(2) end bv x_utf16 = x_utf16.space(1, ',') x_utf8 = x_utf8.space(1, ',') catch ex = UnsupportedEncodingException ex.printStackTrace() end cpName = Character.getName(cp) fmt = - 'CodePoint:' - 'index="'ix.right(3, 0)'"' - 'character_count="'cc'"' - 'id="U+'cp.d2x(5)'"' - 'hex="0x'cp.d2x(6)'"' - 'dec="'icp.right(7, 0)'"' - 'oct="'ocp.right(7, 0)'"' - 'char="'scp'"' - 'utf-16="'x_utf16'"' - 'utf-8="'x_utf8'"' - 'name="'cpName'"' return fmt
-- ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ method runSample(arg) public static
parse arg fileNames if fileNames = then fileNames = 'data/utf8-001.txt' loop while fileNames \= parse fileNames fileName fileNames slurped = readCharacters(fileName) say "Input:" slurped encodingDetails(slurped) end say return
</lang>
- Output:
Input: y䮀𝄞𝄢12 Unicode: length="10" code_point_count="8" string="y䮀𝄞𝄢12" CodePoint: index="000" character_count="1" id="U+00079" hex="0x000079" dec="0000121" oct="0000171" char="y" utf-16="0079" utf-8="79" name="LATIN SMALL LETTER Y" CodePoint: index="001" character_count="1" id="U+000E4" hex="0x0000E4" dec="0000228" oct="0000344" char="ä" utf-16="00E4" utf-8="C3,A4" name="LATIN SMALL LETTER A WITH DIAERESIS" CodePoint: index="002" character_count="1" id="U+000AE" hex="0x0000AE" dec="0000174" oct="0000256" char="®" utf-16="00AE" utf-8="C2,AE" name="REGISTERED SIGN" CodePoint: index="003" character_count="1" id="U+020AC" hex="0x0020AC" dec="0008364" oct="0020254" char="€" utf-16="20AC" utf-8="E2,82,AC" name="EURO SIGN" CodePoint: index="004" character_count="2" id="U+1D11E" hex="0x01D11E" dec="0119070" oct="0350436" char="𝄞" utf-16="D834,DD1E" utf-8="F0,9D,84,9E" name="MUSICAL SYMBOL G CLEF" CodePoint: index="004" character_count="1" id="U+0D834" hex="0x00D834" dec="0055348" oct="0154064" char="?" utf-16="FFFD" utf-8="3F" name="HIGH SURROGATES D834" CodePoint: index="005" character_count="1" id="U+0DD1E" hex="0x00DD1E" dec="0056606" oct="0156436" char="?" utf-16="FFFD" utf-8="3F" name="LOW SURROGATES DD1E" CodePoint: index="006" character_count="2" id="U+1D122" hex="0x01D122" dec="0119074" oct="0350442" char="𝄢" utf-16="D834,DD22" utf-8="F0,9D,84,A2" name="MUSICAL SYMBOL F CLEF" CodePoint: index="006" character_count="1" id="U+0D834" hex="0x00D834" dec="0055348" oct="0154064" char="?" utf-16="FFFD" utf-8="3F" name="HIGH SURROGATES D834" CodePoint: index="007" character_count="1" id="U+0DD22" hex="0x00DD22" dec="0056610" oct="0156442" char="?" utf-16="FFFD" utf-8="3F" name="LOW SURROGATES DD22" CodePoint: index="008" character_count="1" id="U+00031" hex="0x000031" dec="0000049" oct="0000061" char="1" utf-16="0031" utf-8="31" name="DIGIT ONE" CodePoint: index="009" character_count="1" id="U+00032" hex="0x000032" dec="0000050" oct="0000062" char="2" utf-16="0032" utf-8="32" name="DIGIT TWO"
Nim
As most system languages, Nim reads bytes and provides functions to decode bytes into Unicode runes. The normal way to read a stream of UTF-8 characters would be to read the file line by line and decode each line using the “utf-8” iterator which yields UTF-8 characters as strings (one by one) or using the “runes” iterator which yields the UTF-8 characters as Runes (one by one).
As in fact the file would be read line by line, even if the characters are actually yielded one by one, it may be considered as cheating. So, we provide a function and an iterator which read bytes one by one.
<lang Nim>import unicode
proc readUtf8(f: File): string =
## Return next UTF-8 character as a string. while true: result.add f.readChar() if result.validateUtf8() == -1: break
iterator readUtf8(f: File): string =
## Yield successive UTF-8 characters from file "f". var res: string while not f.endOfFile: res.setLen(0) while true: res.add f.readChar() if res.validateUtf8() == -1: break yield res</lang>
Pascal
<lang pascal>(* Read a file char by char *)
program ReadFileByChar; var InputFile,OutputFile: file of char; InputChar: char; begin Assign(InputFile, 'testin.txt'); Reset(InputFile); Assign(OutputFile, 'testout.txt'); Rewrite(OutputFile); while not Eof(InputFile) do begin Read(InputFile, InputChar); Write(OutputFile, InputChar) end; Close(InputFile); Close(OutputFile) end.
</lang>
Perl
<lang perl>binmode STDOUT, ':utf8'; # so we can print wide chars without warning
open my $fh, "<:encoding(UTF-8)", "input.txt" or die "$!\n";
while (read $fh, my $char, 1) {
printf "got character $char [U+%04x]\n", ord $char;
}
close $fh;</lang>
If the contents of the input.txt file are aă€⼥
then the output would be:
got character a [U+0061] got character ă [U+0103] got character € [U+20ac] got character ⼥ [U+2f25]
Phix
Generally I use utf8_to_utf32() on whole lines when I want to do character-counting.
You can find that routine in builtins/utfconv.e, and here is a modified copy that reads precisely one unicode character from a file. If there is a genuine demand for it, I could easily add this to that file permanently, and document/autoinclude it properly.
constant INVALID_UTF8 = #FFFD function get_one_utf8_char(integer fn) -- returns INVALID_UTF8 on error, else a string of 1..4 bytes representing one character object res integer headb, bytes, c -- headb = first byte of utf-8 character: headb = getc(fn) if headb=-1 then return -1 end if res = ""&headb -- calculate length of utf-8 character in bytes (1..4): if headb<0 then bytes = 0 -- (utf-8 starts at #0) elsif headb<=0b01111111 then bytes = 1 -- 0b_0xxx_xxxx elsif headb<=0b10111111 then bytes = 0 -- (it's a tail byte) elsif headb<=0b11011111 then bytes = 2 -- 0b_110x_xxxx elsif headb<=0b11101111 then bytes = 3 -- 0b_1110_xxxx elsif headb<=0b11110100 then bytes = 4 -- 0b_1111_0xzz else bytes = 0 -- (utf-8 ends at #10FFFF) end if -- 2..4 bytes encoding (tail range: 0b_1000_0000..0b_1011_1111); for j=1 to bytes-1 do -- tail bytes are valid? c = getc(fn) if c<#80 or c>#BF then bytes = 0 -- invalid tail byte or eof exit end if res &= c end for -- 1 byte encoding (head range: 0b_0000_0000..0b_0111_1111): if bytes=1 then c = headb -- UTF-8 = ASCII -- 2 bytes encoding (head range: 0b_1100_0000..0b_1101_1111): elsif bytes=2 then c = and_bits(headb, #1F)*#40 + -- 0b110[7..11] headb and_bits(res[2], #3F) -- 0b10[1..6] tail if c>#7FF then ?9/0 end if -- sanity check if c<#80 then -- long form? res = INVALID_UTF8 end if -- 3 bytes encoding (head range: 0b_1110_0000..0b_1110_1111): elsif bytes=3 then c = and_bits(headb, #0F)*#1000 + -- 0b1110[13..16] head and_bits(res[2], #3F)*#40 + -- 0b10[7..12] tail and_bits(res[3], #3F) -- 0b10[1..6] tail if c>#FFFF then ?9/0 end if -- sanity check if c<#800 -- long form? or (c>=#D800 and c<=#DFFF) then -- utf-16 incompatible res = INVALID_UTF8 end if -- 4 bytes encoding (head range: 0b_1111_0000..0b_1111_0111): elsif bytes=4 then c = and_bits(headb, #07)*#040000 + -- 0b11110[19..21] head and_bits(res[2], #3F)*#1000 + -- 0b10[13..18] tail and_bits(res[3], #3F)*#0040 + -- 0b10[7..12] tail and_bits(res[4], #3F) -- 0b10[1..6] tail if c<#10000 -- long form? or c>#10FFFF then res = INVALID_UTF8 -- utf-8 ends at #10FFFF end if -- bytes = 0; current byte is not encoded correctly: else res = INVALID_UTF8 end if return res end function
Test code:
--string utf8 = "aă€⼥" -- (same results as next) string utf8 = utf32_to_utf8({#0061,#0103,#20ac,#2f25}) printf(1,"length of utf8 is %d bytes\n",length(utf8)) integer fn = open("test.txt","wb") puts(fn,utf8) close(fn) fn = open("test.txt","r") for i=1 to 5 do object res = get_one_utf8_char(fn) if string(res) then if platform()=LINUX then printf(1,"char %d (%s) is %d bytes\n",{i,res,length(res)}) else -- unicode and consoles tricky on windows, so I'm -- just avoiding that issue altogther (t)here. printf(1,"char %d is %d bytes\n",{i,length(res)}) end if elsif res=-1 then printf(1,"char %d - EOF\n",i) exit else printf(1,"char %d - INVALID_UTF8\n",i) exit end if end for close(fn)
- Output:
length of utf8 is 9 bytes char 1 is 1 bytes char 2 is 2 bytes char 3 is 3 bytes char 4 is 3 bytes char 5 - EOF
PicoLisp
Pico Lisp uses UTF-8 until told otherwise. <lang PicoLisp> (in "wordlist"
(while (char) (process @))
</lang>
Python
<lang python> def get_next_character(f):
# note: assumes valid utf-8 c = f.read(1) while c: while True: try: yield c.decode('utf-8') except UnicodeDecodeError: # we've encountered a multibyte character # read another byte and try again c += f.read(1) else: # c was a valid char, and was yielded, continue c = f.read(1) break
- Usage:
with open("input.txt","rb") as f:
for c in get_next_character(f): print(c)
</lang>
Python 3 simplifies the handling of text files since you can specify an encoding. <lang python>def get_next_character(f):
"""Reads one character from the given textfile""" c = f.read(1) while c: yield c c = f.read(1)
- Usage:
with open("input.txt", encoding="utf-8") as f:
for c in get_next_character(f): print(c, sep="", end="")</lang>
Racket
Don't we all love self reference? <lang racket>
- lang racket
- This file contains utf-8 charachters
- λ, α, γ ...
(for ([c (in-port read-char (open-input-file "read-file.rkt"))])
(display c))
</lang> Output: <lang racket>
- lang racket
- This file contains utf-8 charachters
- λ, α, γ ...
(for ([c (in-port read-char (open-input-file "read-file.rkt"))])
(display c))
</lang>
Raku
(formerly Perl 6)
Raku has a built in method .getc to get a single character from an open file handle. File handles default to UTF-8, so they will handle multi-byte characters correctly.
To read a single character at a time from the Standard Input terminal; $*IN in Raku: <lang perl6>.say while defined $_ = $*IN.getc;</lang>
Or, from a file: <lang perl6>my $filename = 'whatever';
my $in = open( $filename, :r ) orelse .die;
print $_ while defined $_ = $in.getc;</lang>
REXX
version 1
REXX doesn't support UTF8 encoded wide characters, just bytes.
The task's requirement stated that EOF was to be returned upon reaching the end-of-file, so this programming example was written as a subroutine (procedure).
Note that displaying of characters that may modify screen behavior such as tab usage, backspaces, line feeds, carriage returns, "bells" and others are suppressed, but their hexadecimal equivalents are displayed.
<lang rexx>/*REXX program reads and displays a file char by char, returning 'EOF' when done. */
parse arg iFID . /*iFID: is the fileID to be read. */
/* [↓] show the file's contents. */
if iFID\== then do j=1 until x=='EOF' /*J count's the file's characters. */
x=getchar(iFID); y= /*get a character or an 'EOF'. */ if x>>' ' then y=x /*display X if presentable. */ say right(j, 12) 'character, (hex,char)' c2x(x) y end /*j*/ /* [↑] only display X if not low hex*/
exit /*stick a fork in it, we're all done. */
/*──────────────────────────────────────────────────────────────────────────────────────*/
getchar: procedure; parse arg z; if chars(z)==0 then return 'EOF'; return charin(z)</lang>
input file: ABC
and was created by the DOS command (under Windows/XP): echo 123 [¬ a prime]> ABC
123 [¬ a prime]
output (for the above [ABC] input file:
1 character, (hex,char) 31 1 2 character, (hex,char) 32 2 3 character, (hex,char) 33 3 4 character, (hex,char) 20 5 character, (hex,char) 5B [ 6 character, (hex,char) AA ¬ 7 character, (hex,char) 20 8 character, (hex,char) 61 a 9 character, (hex,char) 20 10 character, (hex,char) 70 p 11 character, (hex,char) 72 r 12 character, (hex,char) 69 i 13 character, (hex,char) 6D m 14 character, (hex,char) 65 e 15 character, (hex,char) 5D ] 16 character, (hex,char) 0D 17 character, (hex,char) 0A 18 character, (hex,char) 454F46 EOF End-Of-File.
version 2
<lang rexx>/* REXX ---------------------------------------------------------------
- 29.12.2013 Walter Pachl
- read one utf8 character at a time
- see http://de.wikipedia.org/wiki/UTF-8#Kodierung
- --------------------------------------------------------------------*/
oid='utf8.txt';'erase' oid /* first create file containing utf8 chars*/ Call charout oid,'79'x Call charout oid,'C3A4'x Call charout oid,'C2AE'x Call charout oid,'E282AC'x Call charout oid,'F09D849E'x Call lineout oid fid='utf8.txt' /* then read it and show the contents */ Do Until c8='EOF'
c8=get_utf8char(fid) Say left(c8,4) c2x(c8) End
Exit
get_utf8char: Procedure
Parse Arg f If chars(f)=0 Then Return 'EOF' c=charin(f) b=c2b(c) If left(b,1)=0 Then Nop Else Do p=pos('0',b) Do i=1 To p-2 If chars(f)=0 Then Do Say 'illegal contents in file' f Leave End c=c||charin(f) End End Return c
c2b: Return x2b(c2x(arg(1)))</lang> output:
y 79 ä C3A4 ® C2AE € E282AC � F09D849E EOF 454F46
Ring
<lang ring> fp = fopen("C:\Ring\ReadMe.txt","r") r = fgetc(fp) while isstring(r)
r = fgetc(fp) see r
end fclose(fp) </lang> Output:
================================================== The Ring Programming Language http://ring-lang.net/ Version 1.0 Release Date : January 25, 2016 Update Date : March 27, 2016 =================================================== Binary release for Microsoft Windows =================================================== Run Start.bat to open Ring Notepad then start learning from the documentation Join Ring Group for questions https://groups.google.com/forum/#!forum/ring-lang Greetings, Mahmoud Fayed msfclipper@yahoo.com http://www.facebook.com/mahmoudfayed1986
Ruby
<lang ruby>File.open('input.txt', 'r:utf-8') do |f|
f.each_char{|c| p c}
end</lang>
or
<lang ruby>File.open('input.txt', 'r:utf-8') do |f|
while c = f.getc p c end
end</lang>
Run BASIC
<lang runbasic>open file.txt" for binary as #f numChars = 1 ' specify number of characters to read a$ = input$(#f,numChars) ' read number of characters specified b$ = input$(#f,1) ' read one character close #f</lang>
Rust
Rust standard library provides hardly any straight-forward way to read single UTF-8 characters from a file. Following code implements an iterator that consumes a byte stream, taking only as many bytes as necessary to decode the next UTF-8 character. It provides quite a complete error report, so that the client code can leverage it to deal with corrupted input.
The decoding code is based on utf8-decode crate originally.
<lang Rust>use std::{
convert::TryFrom, fmt::{Debug, Display, Formatter}, io::Read,
};
pub struct ReadUtf8<I: Iterator> {
source: std::iter::Peekable,
}
impl<R: Read> From<R> for ReadUtf8<std::io::Bytes<R>> {
fn from(source: R) -> Self { ReadUtf8 { source: source.bytes().peekable(), } }
}
impl<I, E> Iterator for ReadUtf8 where
I: Iterator<Item = Result<u8, E>>,
{
type Item = Result<char, Error<E>>;
fn next(&mut self) -> Option<Self::Item> { self.source.next().map(|next| match next { Ok(lead) => self.complete_char(lead), Err(e) => Err(Error::SourceError(e)), }) }
}
impl<I, E> ReadUtf8 where
I: Iterator<Item = Result<u8, E>>,
{
fn continuation(&mut self) -> Result<u32, Error<E>> { if let Some(Ok(byte)) = self.source.peek() { let byte = *byte;
return if byte & 0b1100_0000 == 0b1000_0000 { self.source.next(); Ok((byte & 0b0011_1111) as u32) } else { Err(Error::InvalidByte(byte)) }; }
match self.source.next() { None => Err(Error::InputTruncated), Some(Err(e)) => Err(Error::SourceError(e)), Some(Ok(_)) => unreachable!(), } }
fn complete_char(&mut self, lead: u8) -> Result<char, Error<E>> { let a = lead as u32; // Let's name the bytes in the sequence
let result = if a & 0b1000_0000 == 0 { Ok(a) } else if lead & 0b1110_0000 == 0b1100_0000 { let b = self.continuation()?; Ok((a & 0b0001_1111) << 6 | b) } else if a & 0b1111_0000 == 0b1110_0000 { let b = self.continuation()?; let c = self.continuation()?; Ok((a & 0b0000_1111) << 12 | b << 6 | c) } else if a & 0b1111_1000 == 0b1111_0000 { let b = self.continuation()?; let c = self.continuation()?; let d = self.continuation()?; Ok((a & 0b0000_0111) << 18 | b << 12 | c << 6 | d) } else { Err(Error::InvalidByte(lead)) };
Ok(char::try_from(result?).unwrap()) }
}
- [derive(Debug, Clone)]
pub enum Error<E> {
InvalidByte(u8), InputTruncated, SourceError(E),
}
impl<E: Display> Display for Error<E> {
fn fmt(&self, f: &mut Formatter<'_>) -> std::fmt::Result { match self { Self::InvalidByte(b) => write!(f, "invalid byte 0x{:x}", b), Self::InputTruncated => write!(f, "character truncated"), Self::SourceError(e) => e.fmt(f), } }
}
fn main() -> std::io::Result<()> {
for (index, value) in ReadUtf8::from(std::fs::File::open("test.txt")?).enumerate() { match value { Ok(c) => print!("{}", c),
Err(e) => { print!("\u{fffd}"); eprintln!("offset {}: {}", index, e); } } }
Ok(())
}</lang>
Seed7
The library utf8.s7i
provides the functions openUtf8
and getc.
When a file has been opened with openUtf8
fhe function getc
reads UTF-8 characters from the file.
To allow writing Unicode characters to standard output
the file STD_UTF8_OUT is used.
<lang seed7>$ include "seed7_05.s7i";
include "utf8.s7i";
const proc: main is func
local var file: inFile is STD_NULL; var char: ch is ' '; begin OUT := STD_UTF8_OUT; inFile := openUtf8("readAFileCharacterByCharacterUtf8.in", "r"); if inFile <> STD_NULL then while hasNext(inFile) do ch := getc(inFile); writeln("got character " <& ch <& " [U+" <& ord(ch) radix 16 <& "]"); end while; close(inFile); end if; end func;</lang>
- Output:
When the input file readAFileCharacterByCharacterUtf8.in contains the characters aă€⼥ the output is:
got character a [U+61] got character ă [U+103] got character € [U+20ac] got character ⼥ [U+2f25]
Sidef
<lang ruby>var file = File('input.txt') # the input file contains: "aă€⼥" var fh = file.open_r # equivalent with: file.open('<:utf8') fh.each_char { |char|
printf("got character #{char} [U+%04x]\n", char.ord)
}</lang>
- Output:
got character a [U+0061] got character ă [U+0103] got character € [U+20ac] got character ⼥ [U+2f25]
Smalltalk
<lang smalltalk>|utfStream| utfStream := 'input' asFilename readStream asUTF8EncodedStream. [utfStream atEnd] whileFalse:[
Transcript showCR:'got char ',utfStream next.
]. utfStream close.</lang>
Tcl
To read a single character from a file, use: <lang tcl>set ch [read $channel 1]</lang> This will read multiple bytes sufficient to obtain a Unicode character if a suitable encoding has been configured on the channel. For binary channels, this will always consume exactly one byte. However, the low-level channel buffering logic may consume more than one byte (which only really matters where the channel is being handed on to another process and the channel is over a file descriptor that doesn't support the lseek OS call); the extent of buffering can be controlled via: <lang tcl>fconfigure $channel -buffersize $byteCount</lang> When the channel is only being accessed from Tcl (or via Tcl's C API) it is not normally necessary to adjust this option.
Wren
<lang ecmascript>import "io" for File
File.open("input.txt") { |file|
var offset = 0 var char = "" // stores each byte read till we have a complete UTF encoded character while(true) { var b = file.readBytes(1, offset) if (b == "") return // end of stream char = char + b if (char.codePoints[0] >= 0) { // a UTF encoded character is complete System.write(char) // print it char = "" // reset store } offset = offset + 1 }
}</lang>
zkl
zkl doesn't know much about UTF-8 or Unicode but is able to test whether a string or number is valid UTF-8 or not. This code uses that to build a state machine to decode a byte stream into UTF-8 characters. <lang zkl>fcn readUTF8c(chr,s=""){ // transform UTF-8 character stream
s+=chr; try{ s.len(8); return(s) } catch{ if(s.len()>6) throw(__exception) } // 6 bytes max for UTF-8 return(Void.Again,s); // call me again with s & another character
}</lang> Used to modify a zkl iterator, it can consume any stream-able (files, strings, lists, etc) and provides support for foreach, map, look ahead, push back, etc. <lang zkl>fcn utf8Walker(obj){
obj.walker(3) // read characters .tweak(readUTF8c)
}</lang> <lang zkl>s:="-->\u20AC123"; // --> e2,82,ac,31,32,33 == -->€123 utf8Walker(s).walk().println();
w:=utf8Walker(Data(Void,s,"\n")); // Data is a byte bucket foreach c in (utf8Walker(Data(Void,s,"\n"))){ print(c) }
utf8Walker(Data(Void,0xe2,0x82,"123456")).walk().println(); // € is short 1 byte</lang>
- Output:
L("-","-",">","€","1","2","3") -->€123 VM#1 caught this unhandled exception: ValueError : Invalid UTF-8 string
If you wish to push a UTF-8 stream through one or more functions, you can use the same state machine: <lang zkl>stream:=Data(Void,s,"\n").howza(3); // character stream stream.pump(List,readUTF8c,"print")</lang>
- Output:
-->€123
and returns a list of the eight UTF-8 characters (with newline). Or, if file "foo.txt" contains the characters: <lang zkl>File("foo.txt","rb").howza(3).pump(List,readUTF8c,"print");</lang> produces the same result.
- Programming Tasks
- File handling
- AutoHotkey
- C
- C sharp
- Common Lisp
- Crystal
- Delphi
- System.SysUtils
- System.Classes
- Déjà Vu
- Factor
- FunL
- Go
- Haskell
- J
- Java
- Jq
- Julia
- Kotlin
- Lua
- M2000 Interpreter
- Mathematica
- Wolfram Language
- NetRexx
- Java examples needing attention
- Examples needing attention
- Nim
- Pascal
- Perl
- Phix
- PicoLisp
- Python
- Racket
- Raku
- REXX
- Ring
- Ruby
- Run BASIC
- Rust
- Seed7
- Sidef
- Smalltalk
- Tcl
- Wren
- Zkl
- AWK/Omit
- Brlcad/Omit
- GUISS/Omit
- Openscad/Omit
- TPP/Omit
- Unicode