PostgreSQL: byte sequence 0xc5 0xa0 has no equivalent in LATIN1

Those two bytes are Š in UTF-8, and your database or client encoding is LATIN1, which is true ISO-8859-1 and has no Š. Windows-1252 does have it, so the text looked fine until it reached Postgres. The lasting fix is a UTF8 database. If you're stuck with a single-byte encoding for Czech or Croatian text, WIN1250 covers Š and the other letters those languages need.

What is Š in Serbian Cyrillic?

Ш. Serbian Latin and Cyrillic match letter for letter, so š and ш swap cleanly in both directions. The catch is three Latin digraphs, lj, nj and dž, which each stand for a single Cyrillic letter (љ, њ, џ). A converter that goes one character at a time handles Š fine but breaks those.

What's a lone Š doing in the middle of Italian sentences?

It's the word è. In code page 437 è is byte 138, which Windows-1252 reads as Š, so Questo è vero turns into Questo Š vero. French text breaks the same way, with très becoming trŠs. In Go, decode the file with `charmap.CodePage437.NewDecoder().Bytes(data)` from golang.org/x/text.

ASCII 138

Uppercase S with caronLowercase e with grave accent

ASCII 138 is Š on Windows and è in old DOS.

Š (Capital S with Caron) is byte 138 in Windows-1252, Unicode U+0160. The caron (háček in Czech) turns s into an sh sound in Czech, Slovak, Croatian and Lithuanian, as in Škoda and Šibenik. A useful detail when you're guessing a file's encoding: Š is byte 138 in Windows-1250, the Central European code page, as well. So Czech text read with the wrong one of the two keeps its Š while other letters break. ř, for example, is 0xF8 in Windows-1250, which Windows-1252 shows as ø. In UTF-8, Š is C5 A0, garbled as Å followed by a non-breaking space. Lowercase š is byte 154. The HTML entity is &Scaron;.

è (Small E with Grave) is byte 138 in code page 437, the original IBM PC character set, and maps to Unicode U+00E8. Windows-1252 has Š at 138, so Italian DOS text read as Windows-1252 turns è vero into Š vero and caffè into caffŠ. Code page 850 kept è at 138 too. Code page 437 has no capital È, and code page 850 added one at 212.

Latin LetterLatin Letter

Technical Details

Code Example

</>

#include <stdio.h>

int main(void) {
    /* On Windows (code page 1252) byte 138 is Š.
       Modern terminals expect UTF-8, so print the code point, not the byte. */
    printf("\u0160\n");   /* UTF-8: C5 A0 */

    unsigned char b = 138;         /* the raw Windows-1252 byte */
    printf("%d 0x%02X\n", b, b);    /* 138 0x8A */
    return 0;
}

Code Example

</>

#include <stdio.h>

int main(void) {
    /* Byte 138 is è only on a console using code page 437
       (chcp 437 on Windows, or DOSBox). On a UTF-8 terminal the
       lone byte 0x8A is invalid and prints as garbage, often �. */
    putchar(138);

    /* char is signed on x86, so a plain char holding 0x8A is -118.
       Use unsigned char when you compare or index by byte value. */
    char c = (char)138;
    unsigned char u = 138;
    printf("\n%d %d\n", c, u);   /* -118 138 */

    /* Portable: print the Unicode character as UTF-8 instead. */
    printf("\u00E8\n");          /* C3 A8 */
    return 0;
}

How to Type It

Windows: Hold Alt and type 0138 on the numeric keypad, with Num Lock on. The leading zero picks the Windows code page. No numeric keypad? In Word, type 0160 and press Alt+X.
Mac: Press Control+Command+Space to open the Character Viewer, then search for "latin capital letter s with caron".
Linux: In GNOME, or any desktop running IBus, press Ctrl+Shift+U, type 160, then Space. With a Compose key set up, press Compose, c, S.
Phone: iPhone: tap Shift, long-press S and pick Š. Android (Gboard): tap the character at the top of this page to copy it.

How to Type It

Windows: Hold Alt and type 138 on the numeric keypad, with no leading zero. Windows reads it as a DOS code and inserts è.

è is also ASCII 232 on Windows, which has the full guide.

Frequently Asked Questions

Those two bytes are Š in UTF-8, and your database or client encoding is LATIN1, which is true ISO-8859-1 and has no Š. Windows-1252 does have it, so the text looked fine until it reached Postgres. The lasting fix is a UTF8 database. If you're stuck with a single-byte encoding for Czech or Croatian text, WIN1250 covers Š and the other letters those languages need.
Ш. Serbian Latin and Cyrillic match letter for letter, so š and ш swap cleanly in both directions. The catch is three Latin digraphs, lj, nj and dž, which each stand for a single Cyrillic letter (љ, њ, џ). A converter that goes one character at a time handles Š fine but breaks those.
It's the word è. In code page 437 è is byte 138, which Windows-1252 reads as Š, so Questo è vero turns into Questo Š vero. French text breaks the same way, with très becoming trŠs. In Go, decode the file with `charmap.CodePage437.NewDecoder().Bytes(data)` from golang.org/x/text.

Frequently Asked Questions

It's the word è. In code page 437 è is byte 138, which Windows-1252 reads as Š, so Questo è vero turns into Questo Š vero. French text breaks the same way, with très becoming trŠs. In Go, decode the file with `charmap.CodePage437.NewDecoder().Bytes(data)` from golang.org/x/text.

ASCII 138

Uppercase S with caronLowercase e with grave accent

ASCII 138 is Š on Windows and è in old DOS.

Latin Letter

#include <stdio.h>

int main(void) {

/* On Windows (code page 1252) byte 138 is Š.

Modern terminals expect UTF-8, so print the code point, not the byte. */

printf("\u0160\n"); /* UTF-8: C5 A0 */

unsigned char b = 138; /* the raw Windows-1252 byte */

printf("%d 0x%02X\n", b, b); /* 138 0x8A */

return 0;

}

#include <stdio.h>

int main(void) {

/* Byte 138 is è only on a console using code page 437

(chcp 437 on Windows, or DOSBox). On a UTF-8 terminal the

lone byte 0x8A is invalid and prints as garbage, often �. */

putchar(138);

/* char is signed on x86, so a plain char holding 0x8A is -118.

Use unsigned char when you compare or index by byte value. */

char c = (char)138;

unsigned char u = 138;

printf("\n%d %d\n", c, u); /* -118 138 */

/* Portable: print the Unicode character as UTF-8 instead. */

printf("\u00E8\n"); /* C3 A8 */

return 0;

}