# std::collate<CharT>::transform, do_transform

> C++ Atlas: `cppatlas.com/w/cpp/locale/collate/transform` | Source: https://en.cppreference.com/w/cpp/locale/collate/transform | Domain: Localization

The string transformed so that lexicographic comparison of the transformed strings may be used instead of collating of the originals.

## Official Specification Summary (cppreference)

```cpp
// Defined in header <locale>

public:
string_type transform( const CharT* low, const CharT* high ) const; // (1)

protected:
virtual string_type do_transform( const CharT* low, const CharT* high ) const; // (2)
```

1) Public member function, calls the protected virtual member function `do_transform` of the most derived class.

2) Converts the character sequence `[low,` `high)` to a string that, compared lexicographically (e.g. with `operator<` for strings) with the result of calling `transform()` on another string, produces the same result as calling `do_compare()` on the same two strings.

### Parameters

- `low` - pointer to the first character in the sequence to transform
- `high` - one past the end pointer for the sequence to transform

### Return value

The string transformed so that lexicographic comparison of the transformed strings may be used instead of collating of the originals. In the "C" locale, the returned string is the exact copy of `[low,` `high)`. In other locales, the contents of the returned string are implementation-defined, and the size may be considerably longer.

### Notes

In addition to the use in collation, the implementation-specific format of the transformed string is known to std::regex_traits<>::transform_primary, which is able to extract the equivalence class information.

### Example

```cpp
#include <iomanip>
#include <iostream>
#include <locale>

int main()
{
    std::locale::global(std::locale("sv_SE.utf8"));
    auto& f = std::use_facet<std::collate<wchar_t>>(std::locale());

    std::wstring in1 = L"\u00e4ngel";
    std::wstring in2 = L"\u00e5r";

    std::wstring out1 = f.transform(&in1[0], &in1[0] + in1.size());
    std::wstring out2 = f.transform(&in2[0], &in2[0] + in2.size());

    std::wcout << "In the Swedish locale: ";
    if (out1 < out2)
        std::wcout << in1 << " before " << in2 << '\n';
    else
        std::wcout << in2 << " before " << in1 << '\n';

    std::wcout << "In lexicographic comparison: ";
    if (in1 < in2)
        std::wcout << in1 << " before " << in2 << '\n';
    else
        std::wcout << in2 << " before " << in1 << '\n';
}
```

Output:

```text
In the Swedish locale: år before ängel
In lexicographic comparison: ängel before år
```

### See also

- [`strxfrm`](https://en.cppreference.com/w/cpp/string/byte/strxfrm): transform a string so that `strcmp` would produce the same result as `strcoll` (function)
- [`wcsxfrm`](https://en.cppreference.com/w/cpp/string/wide/wcsxfrm): transform a wide string so that `wcscmp` would produce the same result as `wcscoll` (function)

---
Attribution: content from https://en.cppreference.com/w/cpp/locale/collate/transform (cppreference.com), licensed under CC-BY-SA 3.0 and GFDL. Modified: converted to markdown, study notes added. Offline snapshot of 9 February 2025.