substring from a string
Ruslan Mezentsev
I wonder how to get a substring from a string, such as
from 3 to 7 characters:
"testestestest" -> "test"
Thanks.
Evan Shaw
s1 := "testestestest"
s2 := s1[3:7]
s2 contains "test"
- Evan
Ruslan Mezentsev
package main
import "fmt"
func main() {
str1 := "тестестестест"
str2 := str1[3:7]
fmt.Println(str2)
}
Print:
�с�
On 30 июн, 10:35, Evan Shaw <eds...@gmail.com> wrote:
bflm
> Thanks, but does not work for Russian characters:
str1 := "тестестестест"
fmt.Println(str2)
}
But such mixing of UTF-8 encoding with codepoint access by index is
often probably a wrong way. Either work only with []rune (for random
access), or only with UTF-8 (for sequential access). In the later case
one can also range the UTF-8 to get the runes on by one.
Evan Shaw
> Thanks, but does not work for Russian characters:
Sorry, I should have mentioned that. bflm's solution will work for that case.
- Evan
bflm
func main() {
str1 := "тестестестест"
str2 := string([]int(str1)[3:7])
fmt.Println(str2)
}
Sorry.
CrossWall
Ruslan Mezentsev
Jessta
> Forgive me for asking but what is the rune?
'rune' is just a short name for a unicode code-point. (almost a
character, but some characters are made up of multiple code-points)
In Go it's just an int.
>> But such mixing of UTF-8 encoding with codepoint access by index is
>> often probably a wrong way. Either work only with []rune (for random
>> access), or only with UTF-8 (for sequential access). In the later case
>> one can also range the UTF-8 to get the runes on by one.
--
=====================
http://jessta.id.au
⚛
character-offset into a byte-offset.
Or: http://golang.org/pkg/utf8/#String.Slice
Jan Mercl
Jessta
> Maybe the "strings" package should define a function for converting a
> character-offset into a byte-offset.
>
> Or: http://golang.org/pkg/utf8/#String.Slice
The "strings" package doesn't expect any kind of encoding for the
data. It just deals with strings as immutable []byte.
The "utf8" package doesn't deal with actual characters either, it just
deals with converting utf8 encoded data in to unicode code-points.
The "unicode" package is where this kind of functionality should go
because working out what is actually a 'character' in unicode is more
work than a simple slice.
> On Jun 30, 7:41 am, bflm <befelemepesev...@gmail.com> wrote:
>> Oops, I wrote wrong:
>>
>> func main() {
>> str1 := "тестестестест"
>> str2 := string([]int(str1)[3:7])
>> fmt.Println(str2)
>>
>> }
>>
>> Sorry.
--
=====================
http://jessta.id.au
chris dollin
str1 := "тестестестест"
str2 := str1[3:7]
... is that 3 and 7 are magic numbers. In a real slicing, those
numbers have to come from somewhere sensible, since We
All Know that strings are byte-sequences not character-sequences.
And once the indexes have been found in a rune-respecting
way then the slice just works.
Chris
--
Chris "allusive" Dollin
peterGo
I think that you are missing the point. s[2:5] is slicing the byte
array underlying the string. It's not slicing the UTF-8 encoded
characters forming the string. Therefore, outside the ASCII character
set, your solution doesn't work. For example,
package main
import "fmt"
func main() {
s := "abcdefg"
// byte slice and ASCII character slice
fmt.Println(s[2:5])
s = "αβγδεζη"
fmt.Println(s, len(s), []byte(s))
// byte slice
fmt.Println(s[2:5])
// Unicode code point slice
fmt.Println(string([]int(s)[2:5]))
}
Output:
abcdefg 7 [97 98 99 100 101 102 103]
cde
αβγδεζη 14 [206 177 206 178 206 179 206 180 206 181 206 182 206 183]
β
γδε
Peter
Rob 'Commander' Pike
-rob
Ruslan Mezentsev
str := utf8.NewString("тест本sтестест")
fmt.Println(str.Slice(3,7))
print:
т本sт
⚛
> > Maybe the "strings" package should define a function for converting a
> > character-offset into a byte-offset.
>
> > Or:http://golang.org/pkg/utf8/#String.Slice
'Commander' Pike. He suggested the same thing I did.
> The "strings" package doesn't expect any kind of encoding for the
> data. It just deals with strings as immutable []byte.
> The "utf8" package doesn't deal with actual characters either, it just
> deals with converting utf8 encoded data in to unicode code-points.
Jessta
>> The "strings" package doesn't expect any kind of encoding for the
>> data. It just deals with strings as immutable []byte.
>> The "utf8" package doesn't deal with actual characters either, it just
>> deals with converting utf8 encoded data in to unicode code-points.
>
> Apple is not an apple?
I probably should have said 'glyph' instead of 'character'.
I was referring to the possibility of the use of combining characters
(http://en.wikipedia.org/wiki/Combining_character).
Where a 'glyph' entered by a user or displayed on a screen may be
multiple code-points. Thus slicing on code points without an
understanding of what those code-points mean might miss code-points
that the user expects to be there.
- jessta
--
=====================
http://jessta.id.au