Examples with SliceUtf8.tryGetCodePointAt io.airlift.slice.SliceUtf8.tryGetCodePointAt used on opensource projects

Example 1 with SliceUtf8.tryGetCodePointAt

use of io.airlift.slice.SliceUtf8.tryGetCodePointAt in project presto by prestodb.

the class StringFunctions method hammingDistance.

@Description("computes Hamming distance between two strings")
@ScalarFunction
@LiteralParameters({ "x", "y" })
@SqlType(StandardTypes.BIGINT)
public static long hammingDistance(@SqlType("varchar(x)") Slice left, @SqlType("varchar(y)") Slice right) {
    int distance = 0;
    int leftPosition = 0;
    int rightPosition = 0;
    while (leftPosition < left.length() && rightPosition < right.length()) {
        int codePointLeft = tryGetCodePointAt(left, leftPosition);
        int codePointRight = tryGetCodePointAt(right, rightPosition);
        // the following code treats them as equal if they happen to be of the same length
        if (codePointLeft != codePointRight) {
            distance++;
        }
        leftPosition += codePointLeft >= 0 ? lengthOfCodePoint(codePointLeft) : -codePointLeft;
        rightPosition += codePointRight >= 0 ? lengthOfCodePoint(codePointRight) : -codePointRight;
    }
    checkCondition(leftPosition == left.length() && rightPosition == right.length(), INVALID_FUNCTION_ARGUMENT, "The input strings to hamming_distance function must have the same length");
    return distance;
}

Example 2 with SliceUtf8.tryGetCodePointAt

use of io.airlift.slice.SliceUtf8.tryGetCodePointAt in project presto by prestodb.

the class StringFunctions method pad.

private static Slice pad(Slice text, long targetLength, Slice padString, int paddingOffset) {
    checkCondition(0 <= targetLength && targetLength <= Integer.MAX_VALUE, INVALID_FUNCTION_ARGUMENT, "Target length must be in the range [0.." + Integer.MAX_VALUE + "]");
    checkCondition(padString.length() > 0, INVALID_FUNCTION_ARGUMENT, "Padding string must not be empty");
    int textLength = countCodePoints(text);
    int resultLength = (int) targetLength;
    // if our target length is the same as our string then return our string
    if (textLength == resultLength) {
        return text;
    }
    // if our string is bigger than requested then truncate
    if (textLength > resultLength) {
        return SliceUtf8.substring(text, 0, resultLength);
    }
    // number of bytes in each code point
    int padStringLength = countCodePoints(padString);
    int[] padStringCounts = new int[padStringLength];
    for (int i = 0; i < padStringLength; ++i) {
        padStringCounts[i] = lengthOfCodePointSafe(padString, offsetOfCodePoint(padString, i));
    }
    // preallocate the result
    int bufferSize = text.length();
    for (int i = 0; i < resultLength - textLength; ++i) {
        bufferSize += padStringCounts[i % padStringLength];
    }
    Slice buffer = Slices.allocate(bufferSize);
    // fill in the existing string
    int countBytes = bufferSize - text.length();
    int startPointOfExistingText = (paddingOffset + countBytes) % bufferSize;
    buffer.setBytes(startPointOfExistingText, text);
    // assign the pad string while there's enough space for it
    int byteIndex = paddingOffset;
    for (int i = 0; i < countBytes / padString.length(); ++i) {
        buffer.setBytes(byteIndex, padString);
        byteIndex += padString.length();
    }
    // handle the tail: at most we assign padStringLength - 1 code points
    buffer.setBytes(byteIndex, padString.getBytes(0, paddingOffset + countBytes - byteIndex));
    return buffer;
}

Example 3 with SliceUtf8.tryGetCodePointAt

use of io.airlift.slice.SliceUtf8.tryGetCodePointAt in project presto by prestodb.

the class StringFunctions method fromUtf8.

@Description("decodes the UTF-8 encoded string")
@ScalarFunction
@LiteralParameters("x")
@SqlType(StandardTypes.VARCHAR)
public static Slice fromUtf8(@SqlType(StandardTypes.VARBINARY) Slice slice, @SqlType("varchar(x)") Slice replacementCharacter) {
    int count = countCodePoints(replacementCharacter);
    if (count > 1) {
        throw new PrestoException(INVALID_FUNCTION_ARGUMENT, "Replacement character string must empty or a single character");
    }
    OptionalInt replacementCodePoint;
    if (count == 1) {
        try {
            replacementCodePoint = OptionalInt.of(getCodePointAt(replacementCharacter, 0));
        } catch (InvalidUtf8Exception e) {
            throw new PrestoException(INVALID_FUNCTION_ARGUMENT, "Invalid replacement character");
        }
    } else {
        replacementCodePoint = OptionalInt.empty();
    }
    return SliceUtf8.fixInvalidUtf8(slice, replacementCodePoint);
}

Example 4 with SliceUtf8.tryGetCodePointAt

use of io.airlift.slice.SliceUtf8.tryGetCodePointAt in project presto by prestodb.

the class StringFunctions method safeCountCodePoints.

private static int safeCountCodePoints(Slice slice) {
    int codePoints = 0;
    for (int position = 0; position < slice.length(); ) {
        int codePoint = tryGetCodePointAt(slice, position);
        if (codePoint < 0) {
            throw new PrestoException(INVALID_FUNCTION_ARGUMENT, "Invalid UTF-8 encoding in characters: " + slice.toStringUtf8());
        }
        position += lengthOfCodePoint(codePoint);
        codePoints++;
    }
    return codePoints;
}

Example 5 with SliceUtf8.tryGetCodePointAt

use of io.airlift.slice.SliceUtf8.tryGetCodePointAt in project presto by prestodb.

the class CharacterStringCasts method codePointsToSliceUtf8.

private static Slice codePointsToSliceUtf8(List<Integer> codePoints) {
    int length = codePoints.stream().mapToInt(SliceUtf8::lengthOfCodePoint).sum();
    Slice result = Slices.wrappedBuffer(new byte[length]);
    int offset = 0;
    for (int codePoint : codePoints) {
        setCodePointAt(codePoint, result, offset);
        offset += lengthOfCodePoint(codePoint);
    }
    return result;
}