java - JDBC:如何读取MySQL点格式

Question

我在 MySQL 中有一个 2D POINT 列。我使用 JDBC 将其读入 Java 对象。结果是一个字节[25]。我想知道这是什么格式，以便我可以将其转换为我的观点(例如 double 值)

点是这样设置的:

UPDATE ... SET point=GeomFromText('POINT(18 63)')

数据库中的结果字节:

0x00000000010100000000000000000032400000000000804F40

Answer 1

因此，对于解决方案，我使用了 WKB 阅读器中的代码:https://github.com/simplegeo/jts/tree/master/src/com/vividsolutions/jts/io

因为我不想包含所有内容，所以我只使用了从 ByteOrderValues.java 中以字节序格式读取 double 的代码:

/**
     * Read a long from a byte buffer in big endian format.
     * @param buf must be 8 bytes
     */
    public static long readLongFromBytesBigEndian(byte[] buf) {
            return    (long) (buf[0] & 0xff) << 56
                    | (long) (buf[1] & 0xff) << 48
                    | (long) (buf[2] & 0xff) << 40
                    | (long) (buf[3] & 0xff) << 32
                    | (long) (buf[4] & 0xff) << 24
                    | (long) (buf[5] & 0xff) << 16
                    | (long) (buf[6] & 0xff) <<  8
                    | (long) (buf[7] & 0xff);
    }

    /**
     * Read a long from a byte buffer in little endian format.
     * @param buf must be 8 bytes
     */
    public static long readLongFromBytesLittleEndian(byte[] buf) {
            return    (long) (buf[7] & 0xff) << 56
                    | (long) (buf[6] & 0xff) << 48
                    | (long) (buf[5] & 0xff) << 40
                    | (long) (buf[4] & 0xff) << 32
                    | (long) (buf[3] & 0xff) << 24
                    | (long) (buf[2] & 0xff) << 16
                    | (long) (buf[1] & 0xff) <<  8
                    | (long) (buf[0] & 0xff);
    }

    /**
     * Read a long from a byte buffer in big or little endian format.
     * @param bigEndian true for big endian, false for little endian
     * @param buf must be 8 bytes
     */
    public static double readDoubleFromBytes(byte[] buf, boolean bigEndian) {
        long longVal =  bigEndian ? readLongFromBytesBigEndian(buf)
                : readLongFromBytesLittleEndian(buf);
        return Double.longBitsToDouble(longVal);
    }

    /**
     * Read a long from a byte buffer in big or little endian format.
     * @param bigEndian true for big endian, false for little endian
     * @param buf must be 8 bytes length after offset
     */
    public static double readDoubleFromBytes(byte[] buf, int offset, boolean bigEndian) {
        byte[] bufOf8Bytes = Arrays.copyOfRange(buf, offset, offset + 8);
        return readDoubleFromBytes(bufOf8Bytes, bigEndian);
    }

    /**
     * Read a coordinate from a byte array in WKB format.
     * @param wkbBytes must be 25 bytes long
     */
    public static Coordinate readCoordinateFromWkbBytes(byte[] wkbBytes) {
        // Points are stored in MySQL marked as big endian, but in reality is little endian. Not good
        boolean isBigEndian = false; // readIsWkbBigEndianByteOrder(wkbBytes[0]);
        double x = readDoubleFromBytes(wkbBytes, 9, isBigEndian);
        double y = readDoubleFromBytes(wkbBytes, 17, isBigEndian);
        Coordinate coordinate = new Coordinate();
        coordinate.setX(x);
        coordinate.setY(y);
        return coordinate;
    }

    public static boolean readIsWkbBigEndianByteOrder(byte b) {
        final byte BIG_ENDIAN = 0;
        final byte LITTLE_ENDIAN = 1;
        return b == BIG_ENDIAN;
    }

注意:我点的第一个字节，表示big/little endian，设置为0，根据WKBReader应该是big endian。但是我发现 littl endian 才是实际的格式。