varbit.c

PostgreSQL 8.1.4的源码 适用于Linux下的开源数据库系统

	bits8	   *p1,
			   *p2,
			   *r;
	bits8		mask;

	bitlen1 = VARBITLEN(arg1);
	bitlen2 = VARBITLEN(arg2);
	if (bitlen1 != bitlen2)
		ereport(ERROR,
				(errcode(ERRCODE_STRING_DATA_LENGTH_MISMATCH),
				 errmsg("cannot OR bit strings of different sizes")));
	len = VARSIZE(arg1);
	result = (VarBit *) palloc(len);
	VARATT_SIZEP(result) = len;
	VARBITLEN(result) = bitlen1;

	p1 = VARBITS(arg1);
	p2 = VARBITS(arg2);
	r = VARBITS(result);
	for (i = 0; i < VARBITBYTES(arg1); i++)
		*r++ = *p1++ | *p2++;

	/* Pad the result */
	mask = BITMASK << VARBITPAD(result);
	if (mask)
	{
		r--;
		*r &= mask;
	}

	PG_RETURN_VARBIT_P(result);
}

/* bitxor
 * perform a logical XOR on two bit strings.
 */
Datum
bitxor(PG_FUNCTION_ARGS)
{
	VarBit	   *arg1 = PG_GETARG_VARBIT_P(0);
	VarBit	   *arg2 = PG_GETARG_VARBIT_P(1);
	VarBit	   *result;
	int			len,
				bitlen1,
				bitlen2,
				i;
	bits8	   *p1,
			   *p2,
			   *r;
	bits8		mask;

	bitlen1 = VARBITLEN(arg1);
	bitlen2 = VARBITLEN(arg2);
	if (bitlen1 != bitlen2)
		ereport(ERROR,
				(errcode(ERRCODE_STRING_DATA_LENGTH_MISMATCH),
				 errmsg("cannot XOR bit strings of different sizes")));

	len = VARSIZE(arg1);
	result = (VarBit *) palloc(len);
	VARATT_SIZEP(result) = len;
	VARBITLEN(result) = bitlen1;

	p1 = VARBITS(arg1);
	p2 = VARBITS(arg2);
	r = VARBITS(result);
	for (i = 0; i < VARBITBYTES(arg1); i++)
		*r++ = *p1++ ^ *p2++;

	/* Pad the result */
	mask = BITMASK << VARBITPAD(result);
	if (mask)
	{
		r--;
		*r &= mask;
	}

	PG_RETURN_VARBIT_P(result);
}

/* bitnot
 * perform a logical NOT on a bit string.
 */
Datum
bitnot(PG_FUNCTION_ARGS)
{
	VarBit	   *arg = PG_GETARG_VARBIT_P(0);
	VarBit	   *result;
	bits8	   *p,
			   *r;
	bits8		mask;

	result = (VarBit *) palloc(VARSIZE(arg));
	VARATT_SIZEP(result) = VARSIZE(arg);
	VARBITLEN(result) = VARBITLEN(arg);

	p = VARBITS(arg);
	r = VARBITS(result);
	for (; p < VARBITEND(arg); p++)
		*r++ = ~*p;

	/* Pad the result */
	mask = BITMASK << VARBITPAD(result);
	if (mask)
	{
		r--;
		*r &= mask;
	}

	PG_RETURN_VARBIT_P(result);
}

/* bitshiftleft
 * do a left shift (i.e. towards the beginning of the string)
 */
Datum
bitshiftleft(PG_FUNCTION_ARGS)
{
	VarBit	   *arg = PG_GETARG_VARBIT_P(0);
	int32		shft = PG_GETARG_INT32(1);
	VarBit	   *result;
	int			byte_shift,
				ishift,
				len;
	bits8	   *p,
			   *r;

	/* Negative shift is a shift to the right */
	if (shft < 0)
		PG_RETURN_DATUM(DirectFunctionCall2(bitshiftright,
											VarBitPGetDatum(arg),
											Int32GetDatum(-shft)));

	result = (VarBit *) palloc(VARSIZE(arg));
	VARATT_SIZEP(result) = VARSIZE(arg);
	VARBITLEN(result) = VARBITLEN(arg);
	r = VARBITS(result);

	/* If we shifted all the bits out, return an all-zero string */
	if (shft >= VARBITLEN(arg))
	{
		MemSet(r, 0, VARBITBYTES(arg));
		PG_RETURN_VARBIT_P(result);
	}

	byte_shift = shft / BITS_PER_BYTE;
	ishift = shft % BITS_PER_BYTE;
	p = VARBITS(arg) + byte_shift;

	if (ishift == 0)
	{
		/* Special case: we can do a memcpy */
		len = VARBITBYTES(arg) - byte_shift;
		memcpy(r, p, len);
		MemSet(r + len, 0, byte_shift);
	}
	else
	{
		for (; p < VARBITEND(arg); r++)
		{
			*r = *p << ishift;
			if ((++p) < VARBITEND(arg))
				*r |= *p >> (BITS_PER_BYTE - ishift);
		}
		for (; r < VARBITEND(result); r++)
			*r = 0;
	}

	PG_RETURN_VARBIT_P(result);
}

/* bitshiftright
 * do a right shift (i.e. towards the end of the string)
 */
Datum
bitshiftright(PG_FUNCTION_ARGS)
{
	VarBit	   *arg = PG_GETARG_VARBIT_P(0);
	int32		shft = PG_GETARG_INT32(1);
	VarBit	   *result;
	int			byte_shift,
				ishift,
				len;
	bits8	   *p,
			   *r;

	/* Negative shift is a shift to the left */
	if (shft < 0)
		PG_RETURN_DATUM(DirectFunctionCall2(bitshiftleft,
											VarBitPGetDatum(arg),
											Int32GetDatum(-shft)));

	result = (VarBit *) palloc(VARSIZE(arg));
	VARATT_SIZEP(result) = VARSIZE(arg);
	VARBITLEN(result) = VARBITLEN(arg);
	r = VARBITS(result);

	/* If we shifted all the bits out, return an all-zero string */
	if (shft >= VARBITLEN(arg))
	{
		MemSet(r, 0, VARBITBYTES(arg));
		PG_RETURN_VARBIT_P(result);
	}

	byte_shift = shft / BITS_PER_BYTE;
	ishift = shft % BITS_PER_BYTE;
	p = VARBITS(arg);

	/* Set the first part of the result to 0 */
	MemSet(r, 0, byte_shift);
	r += byte_shift;

	if (ishift == 0)
	{
		/* Special case: we can do a memcpy */
		len = VARBITBYTES(arg) - byte_shift;
		memcpy(r, p, len);
	}
	else
	{
		if (r < VARBITEND(result))
			*r = 0;				/* initialize first byte */
		for (; r < VARBITEND(result); p++)
		{
			*r |= *p >> ishift;
			if ((++r) < VARBITEND(result))
				*r = (*p << (BITS_PER_BYTE - ishift)) & BITMASK;
		}
	}

	PG_RETURN_VARBIT_P(result);
}

/*
 * This is not defined in any standard. We retain the natural ordering of
 * bits here, as it just seems more intuitive.
 */
Datum
bitfromint4(PG_FUNCTION_ARGS)
{
	int32		a = PG_GETARG_INT32(0);
	int32		typmod = PG_GETARG_INT32(1);
	VarBit	   *result;
	bits8	   *r;
	int			rlen;
	int			destbitsleft,
				srcbitsleft;

	if (typmod <= 0)
		typmod = 1;				/* default bit length */

	rlen = VARBITTOTALLEN(typmod);
	result = (VarBit *) palloc(rlen);
	VARATT_SIZEP(result) = rlen;
	VARBITLEN(result) = typmod;

	r = VARBITS(result);
	destbitsleft = typmod;
	srcbitsleft = 32;
	/* drop any input bits that don't fit */
	srcbitsleft = Min(srcbitsleft, destbitsleft);
	/* sign-fill any excess bytes in output */
	while (destbitsleft >= srcbitsleft + 8)
	{
		*r++ = (bits8) ((a < 0) ? BITMASK : 0);
		destbitsleft -= 8;
	}
	/* store first fractional byte */
	if (destbitsleft > srcbitsleft)
	{
		*r++ = (bits8) ((a >> (srcbitsleft - 8)) & BITMASK);
		destbitsleft -= 8;
	}
	/* Now srcbitsleft and destbitsleft are the same, need not track both */
	/* store whole bytes */
	while (destbitsleft >= 8)
	{
		*r++ = (bits8) ((a >> (destbitsleft - 8)) & BITMASK);
		destbitsleft -= 8;
	}
	/* store last fractional byte */
	if (destbitsleft > 0)
		*r = (bits8) ((a << (8 - destbitsleft)) & BITMASK);

	PG_RETURN_VARBIT_P(result);
}

Datum
bittoint4(PG_FUNCTION_ARGS)
{
	VarBit	   *arg = PG_GETARG_VARBIT_P(0);
	uint32		result;
	bits8	   *r;

	/* Check that the bit string is not too long */
	if (VARBITLEN(arg) > sizeof(result) * BITS_PER_BYTE)
		ereport(ERROR,
				(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
				 errmsg("integer out of range")));

	result = 0;
	for (r = VARBITS(arg); r < VARBITEND(arg); r++)
	{
		result <<= BITS_PER_BYTE;
		result |= *r;
	}
	/* Now shift the result to take account of the padding at the end */
	result >>= VARBITPAD(arg);

	PG_RETURN_INT32(result);
}

Datum
bitfromint8(PG_FUNCTION_ARGS)
{
	int64		a = PG_GETARG_INT64(0);
	int32		typmod = PG_GETARG_INT32(1);
	VarBit	   *result;
	bits8	   *r;
	int			rlen;
	int			destbitsleft,
				srcbitsleft;

	if (typmod <= 0)
		typmod = 1;				/* default bit length */

	rlen = VARBITTOTALLEN(typmod);
	result = (VarBit *) palloc(rlen);
	VARATT_SIZEP(result) = rlen;
	VARBITLEN(result) = typmod;

	r = VARBITS(result);
	destbitsleft = typmod;
#ifndef INT64_IS_BUSTED
	srcbitsleft = 64;
#else
	srcbitsleft = 32;			/* don't try to shift more than 32 */
#endif
	/* drop any input bits that don't fit */
	srcbitsleft = Min(srcbitsleft, destbitsleft);
	/* sign-fill any excess bytes in output */
	while (destbitsleft >= srcbitsleft + 8)
	{
		*r++ = (bits8) ((a < 0) ? BITMASK : 0);
		destbitsleft -= 8;
	}
	/* store first fractional byte */
	if (destbitsleft > srcbitsleft)
	{
		*r++ = (bits8) ((a >> (srcbitsleft - 8)) & BITMASK);
		destbitsleft -= 8;
	}
	/* Now srcbitsleft and destbitsleft are the same, need not track both */
	/* store whole bytes */
	while (destbitsleft >= 8)
	{
		*r++ = (bits8) ((a >> (destbitsleft - 8)) & BITMASK);
		destbitsleft -= 8;
	}
	/* store last fractional byte */
	if (destbitsleft > 0)
		*r = (bits8) ((a << (8 - destbitsleft)) & BITMASK);

	PG_RETURN_VARBIT_P(result);
}

Datum
bittoint8(PG_FUNCTION_ARGS)
{
	VarBit	   *arg = PG_GETARG_VARBIT_P(0);
	uint64		result;
	bits8	   *r;

	/* Check that the bit string is not too long */
	if (VARBITLEN(arg) > sizeof(result) * BITS_PER_BYTE)
		ereport(ERROR,
				(errcode(ERRCODE_NUMERIC_VALUE_OUT_OF_RANGE),
				 errmsg("bigint out of range")));

	result = 0;
	for (r = VARBITS(arg); r < VARBITEND(arg); r++)
	{
		result <<= BITS_PER_BYTE;
		result |= *r;
	}
	/* Now shift the result to take account of the padding at the end */
	result >>= VARBITPAD(arg);

	PG_RETURN_INT64(result);
}


/* Determines the position of S2 in the bitstring S1 (1-based string).
 * If S2 does not appear in S1 this function returns 0.
 * If S2 is of length 0 this function returns 1.
 * Compatible in usage with POSITION() functions for other data types.
 */
Datum
bitposition(PG_FUNCTION_ARGS)
{
	VarBit	   *str = PG_GETARG_VARBIT_P(0);
	VarBit	   *substr = PG_GETARG_VARBIT_P(1);
	int			substr_length,
				str_length,
				i,
				is;
	bits8	   *s,				/* pointer into substring */
			   *p;				/* pointer into str */
	bits8		cmp,			/* shifted substring byte to compare */
				mask1,			/* mask for substring byte shifted right */
				mask2,			/* mask for substring byte shifted left */
				end_mask,		/* pad mask for last substring byte */
				str_mask;		/* pad mask for last string byte */
	bool		is_match;

	/* Get the substring length */
	substr_length = VARBITLEN(substr);
	str_length = VARBITLEN(str);

	/* String has zero length or substring longer than string, return 0 */
	if ((str_length == 0) || (substr_length > str_length))
		PG_RETURN_INT32(0);

	/* zero-length substring means return 1 */
	if (substr_length == 0)
		PG_RETURN_INT32(1);

	/* Initialise the padding masks */
	end_mask = BITMASK << VARBITPAD(substr);
	str_mask = BITMASK << VARBITPAD(str);
	for (i = 0; i < VARBITBYTES(str) - VARBITBYTES(substr) + 1; i++)
	{
		for (is = 0; is < BITS_PER_BYTE; is++)
		{
			is_match = true;
			p = VARBITS(str) + i;
			mask1 = BITMASK >> is;
			mask2 = ~mask1;
			for (s = VARBITS(substr);
				 is_match && s < VARBITEND(substr); s++)
			{
				cmp = *s >> is;
				if (s == VARBITEND(substr) - 1)
				{
					mask1 &= end_mask >> is;
					if (p == VARBITEND(str) - 1)
					{
						/* Check that there is enough of str left */
						if (mask1 & ~str_mask)
						{
							is_match = false;
							break;
						}
						mask1 &= str_mask;
					}
				}
				is_match = ((cmp ^ *p) & mask1) == 0;
				if (!is_match)
					break;
				/* Move on to the next byte */
				p++;
				if (p == VARBITEND(str))
				{
					mask2 = end_mask << (BITS_PER_BYTE - is);
					is_match = mask2 == 0;
#if 0
					elog(DEBUG4, "S. %d %d em=%2x sm=%2x r=%d",
						 i, is, end_mask, mask2, is_match);
#endif
					break;
				}
				cmp = *s << (BITS_PER_BYTE - is);
				if (s == VARBITEND(substr) - 1)
				{
					mask2 &= end_mask << (BITS_PER_BYTE - is);
					if (p == VARBITEND(str) - 1)
					{
						if (mask2 & ~str_mask)
						{
							is_match = false;
							break;
						}
						mask2 &= str_mask;
					}
				}
				is_match = ((cmp ^ *p) & mask2) == 0;
			}
			/* Have we found a match? */
			if (is_match)
				PG_RETURN_INT32(i * BITS_PER_BYTE + is + 1);
		}
	}
	PG_RETURN_INT32(0);
}