varlena.c

postgresql8.3.4源码,开源数据库

		{
			/*
			 * If escape char is not followed by any expected char, just treat
			 * it as ordinary data to copy.  (XXX would it be better to throw
			 * an error?)
			 */
			appendStringInfoChar(str, '\\');
			continue;
		}

		if (so != -1 && eo != -1)
		{
			/*
			 * Copy the text that is back reference of regexp.	Note so and eo
			 * are counted in characters not bytes.
			 */
			char	   *chunk_start;
			int			chunk_len;

			Assert(so >= data_pos);
			chunk_start = start_ptr;
			chunk_start += charlen_to_bytelen(chunk_start, so - data_pos);
			chunk_len = charlen_to_bytelen(chunk_start, eo - so);
			appendBinaryStringInfo(str, chunk_start, chunk_len);
		}
	}
}

#define REGEXP_REPLACE_BACKREF_CNT		10

/*
 * replace_text_regexp
 *
 * replace text that matches to regexp in src_text to replace_text.
 *
 * Note: to avoid having to include regex.h in builtins.h, we declare
 * the regexp argument as void *, but really it's regex_t *.
 */
text *
replace_text_regexp(text *src_text, void *regexp,
					text *replace_text, bool glob)
{
	text	   *ret_text;
	regex_t    *re = (regex_t *) regexp;
	int			src_text_len = VARSIZE_ANY_EXHDR(src_text);
	StringInfoData buf;
	regmatch_t	pmatch[REGEXP_REPLACE_BACKREF_CNT];
	pg_wchar   *data;
	size_t		data_len;
	int			search_start;
	int			data_pos;
	char	   *start_ptr;
	bool		have_escape;

	initStringInfo(&buf);

	/* Convert data string to wide characters. */
	data = (pg_wchar *) palloc((src_text_len + 1) * sizeof(pg_wchar));
	data_len = pg_mb2wchar_with_len(VARDATA_ANY(src_text), data, src_text_len);

	/* Check whether replace_text has escape char. */
	have_escape = check_replace_text_has_escape_char(replace_text);

	/* start_ptr points to the data_pos'th character of src_text */
	start_ptr = (char *) VARDATA_ANY(src_text);
	data_pos = 0;

	search_start = 0;
	while (search_start <= data_len)
	{
		int			regexec_result;

		CHECK_FOR_INTERRUPTS();

		regexec_result = pg_regexec(re,
									data,
									data_len,
									search_start,
									NULL,		/* no details */
									REGEXP_REPLACE_BACKREF_CNT,
									pmatch,
									0);

		if (regexec_result == REG_NOMATCH)
			break;

		if (regexec_result != REG_OKAY)
		{
			char		errMsg[100];

			pg_regerror(regexec_result, re, errMsg, sizeof(errMsg));
			ereport(ERROR,
					(errcode(ERRCODE_INVALID_REGULAR_EXPRESSION),
					 errmsg("regular expression failed: %s", errMsg)));
		}

		/*
		 * Copy the text to the left of the match position.  Note we are given
		 * character not byte indexes.
		 */
		if (pmatch[0].rm_so - data_pos > 0)
		{
			int			chunk_len;

			chunk_len = charlen_to_bytelen(start_ptr,
										   pmatch[0].rm_so - data_pos);
			appendBinaryStringInfo(&buf, start_ptr, chunk_len);

			/*
			 * Advance start_ptr over that text, to avoid multiple rescans of
			 * it if the replace_text contains multiple back-references.
			 */
			start_ptr += chunk_len;
			data_pos = pmatch[0].rm_so;
		}

		/*
		 * Copy the replace_text. Process back references when the
		 * replace_text has escape characters.
		 */
		if (have_escape)
			appendStringInfoRegexpSubstr(&buf, replace_text, pmatch,
										 start_ptr, data_pos);
		else
			appendStringInfoText(&buf, replace_text);

		/* Advance start_ptr and data_pos over the matched text. */
		start_ptr += charlen_to_bytelen(start_ptr,
										pmatch[0].rm_eo - data_pos);
		data_pos = pmatch[0].rm_eo;

		/*
		 * When global option is off, replace the first instance only.
		 */
		if (!glob)
			break;

		/*
		 * Search from next character when the matching text is zero width.
		 */
		search_start = data_pos;
		if (pmatch[0].rm_so == pmatch[0].rm_eo)
			search_start++;
	}

	/*
	 * Copy the text to the right of the last match.
	 */
	if (data_pos < data_len)
	{
		int			chunk_len;

		chunk_len = ((char *) src_text + VARSIZE_ANY(src_text)) - start_ptr;
		appendBinaryStringInfo(&buf, start_ptr, chunk_len);
	}

	ret_text = PG_STR_GET_TEXT(buf.data);
	pfree(buf.data);
	pfree(data);

	return ret_text;
}

/*
 * split_text
 * parse input string
 * return ord item (1 based)
 * based on provided field separator
 */
Datum
split_text(PG_FUNCTION_ARGS)
{
	text	   *inputstring = PG_GETARG_TEXT_PP(0);
	text	   *fldsep = PG_GETARG_TEXT_PP(1);
	int			fldnum = PG_GETARG_INT32(2);
	int			inputstring_len;
	int			fldsep_len;
	TextPositionState state;
	int			start_posn;
	int			end_posn;
	text	   *result_text;

	/* field number is 1 based */
	if (fldnum < 1)
		ereport(ERROR,
				(errcode(ERRCODE_INVALID_PARAMETER_VALUE),
				 errmsg("field position must be greater than zero")));

	text_position_setup(inputstring, fldsep, &state);

	/*
	 * Note: we check the converted string length, not the original, because
	 * they could be different if the input contained invalid encoding.
	 */
	inputstring_len = state.len1;
	fldsep_len = state.len2;

	/* return empty string for empty input string */
	if (inputstring_len < 1)
	{
		text_position_cleanup(&state);
		PG_RETURN_TEXT_P(PG_STR_GET_TEXT(""));
	}

	/* empty field separator */
	if (fldsep_len < 1)
	{
		text_position_cleanup(&state);
		/* if first field, return input string, else empty string */
		if (fldnum == 1)
			PG_RETURN_TEXT_P(inputstring);
		else
			PG_RETURN_TEXT_P(PG_STR_GET_TEXT(""));
	}

	/* identify bounds of first field */
	start_posn = 1;
	end_posn = text_position_next(1, &state);

	/* special case if fldsep not found at all */
	if (end_posn == 0)
	{
		text_position_cleanup(&state);
		/* if field 1 requested, return input string, else empty string */
		if (fldnum == 1)
			PG_RETURN_TEXT_P(inputstring);
		else
			PG_RETURN_TEXT_P(PG_STR_GET_TEXT(""));
	}

	while (end_posn > 0 && --fldnum > 0)
	{
		/* identify bounds of next field */
		start_posn = end_posn + fldsep_len;
		end_posn = text_position_next(start_posn, &state);
	}

	text_position_cleanup(&state);

	if (fldnum > 0)
	{
		/* N'th field separator not found */
		/* if last field requested, return it, else empty string */
		if (fldnum == 1)
			result_text = text_substring(PointerGetDatum(inputstring),
										 start_posn,
										 -1,
										 true);
		else
			result_text = PG_STR_GET_TEXT("");
	}
	else
	{
		/* non-last field requested */
		result_text = text_substring(PointerGetDatum(inputstring),
									 start_posn,
									 end_posn - start_posn,
									 false);
	}

	PG_RETURN_TEXT_P(result_text);
}

/*
 * text_to_array
 * parse input string
 * return text array of elements
 * based on provided field separator
 */
Datum
text_to_array(PG_FUNCTION_ARGS)
{
	text	   *inputstring = PG_GETARG_TEXT_PP(0);
	text	   *fldsep = PG_GETARG_TEXT_PP(1);
	int			inputstring_len;
	int			fldsep_len;
	TextPositionState state;
	int			fldnum;
	int			start_posn;
	int			end_posn;
	int			chunk_len;
	char	   *start_ptr;
	text	   *result_text;
	ArrayBuildState *astate = NULL;

	text_position_setup(inputstring, fldsep, &state);

	/*
	 * Note: we check the converted string length, not the original, because
	 * they could be different if the input contained invalid encoding.
	 */
	inputstring_len = state.len1;
	fldsep_len = state.len2;

	/* return NULL for empty input string */
	if (inputstring_len < 1)
	{
		text_position_cleanup(&state);
		PG_RETURN_NULL();
	}

	/*
	 * empty field separator return one element, 1D, array using the input
	 * string
	 */
	if (fldsep_len < 1)
	{
		text_position_cleanup(&state);
		PG_RETURN_ARRAYTYPE_P(create_singleton_array(fcinfo, TEXTOID,
										   PointerGetDatum(inputstring), 1));
	}

	start_posn = 1;
	/* start_ptr points to the start_posn'th character of inputstring */
	start_ptr = VARDATA_ANY(inputstring);

	for (fldnum = 1;; fldnum++) /* field number is 1 based */
	{
		CHECK_FOR_INTERRUPTS();

		end_posn = text_position_next(start_posn, &state);

		if (end_posn == 0)
		{
			/* fetch last field */
			chunk_len = ((char *) inputstring + VARSIZE_ANY(inputstring)) - start_ptr;
		}
		else
		{
			/* fetch non-last field */
			chunk_len = charlen_to_bytelen(start_ptr, end_posn - start_posn);
		}

		/* must build a temp text datum to pass to accumArrayResult */
		result_text = (text *) palloc(VARHDRSZ + chunk_len);
		SET_VARSIZE(result_text, VARHDRSZ + chunk_len);
		memcpy(VARDATA(result_text), start_ptr, chunk_len);

		/* stash away this field */
		astate = accumArrayResult(astate,
								  PointerGetDatum(result_text),
								  false,
								  TEXTOID,
								  CurrentMemoryContext);

		pfree(result_text);

		if (end_posn == 0)
			break;

		start_posn = end_posn;
		start_ptr += chunk_len;
		start_posn += fldsep_len;
		start_ptr += charlen_to_bytelen(start_ptr, fldsep_len);
	}

	text_position_cleanup(&state);

	PG_RETURN_ARRAYTYPE_P(makeArrayResult(astate,
										  CurrentMemoryContext));
}

/*
 * array_to_text
 * concatenate Cstring representation of input array elements
 * using provided field separator
 */
Datum
array_to_text(PG_FUNCTION_ARGS)
{
	ArrayType  *v = PG_GETARG_ARRAYTYPE_P(0);
	char	   *fldsep = PG_TEXTARG_GET_STR(1);
	int			nitems,
			   *dims,
				ndims;
	Oid			element_type;
	int			typlen;
	bool		typbyval;
	char		typalign;
	StringInfoData buf;
	bool		printed = false;
	char	   *p;
	bits8	   *bitmap;
	int			bitmask;
	int			i;
	ArrayMetaState *my_extra;

	ndims = ARR_NDIM(v);
	dims = ARR_DIMS(v);
	nitems = ArrayGetNItems(ndims, dims);

	/* if there are no elements, return an empty string */
	if (nitems == 0)
		PG_RETURN_TEXT_P(PG_STR_GET_TEXT(""));

	element_type = ARR_ELEMTYPE(v);
	initStringInfo(&buf);

	/*
	 * We arrange to look up info about element type, including its output
	 * conversion proc, only once per series of calls, assuming the element
	 * type doesn't change underneath us.
	 */
	my_extra = (ArrayMetaState *) fcinfo->flinfo->fn_extra;
	if (my_extra == NULL)
	{
		fcinfo->flinfo->fn_extra = MemoryContextAlloc(fcinfo->flinfo->fn_mcxt,
													  sizeof(ArrayMetaState));
		my_extra = (ArrayMetaState *) fcinfo->flinfo->fn_extra;
		my_extra->element_type = ~element_type;
	}

	if (my_extra->element_type != element_type)
	{
		/*
		 * Get info about element type, including its output conversion proc
		 */
		get_type_io_data(element_type, IOFunc_output,
						 &my_extra->typlen, &my_extra->typbyval,
						 &my_extra->typalign, &my_extra->typdelim,
						 &my_extra->typioparam, &my_extra->typiofunc);
		fmgr_info_cxt(my_extra->typiofunc, &my_extra->proc,
					  fcinfo->flinfo->fn_mcxt);
		my_extra->element_type = element_type;
	}
	typlen = my_extra->typlen;
	typbyval = my_extra->typbyval;
	typalign = my_extra->typalign;

	p = ARR_DATA_PTR(v);
	bitmap = ARR_NULLBITMAP(v);
	bitmask = 1;

	for (i = 0; i < nitems; i++)
	{
		Datum		itemvalue;
		char	   *value;

		/* Get source element, checking for NULL */
		if (bitmap && (*bitmap & bitmask) == 0)
		{
			/* we ignore nulls */
		}
		else
		{
			itemvalue = fetch_att(p, typbyval, typlen);

			value = OutputFunctionCall(&my_extra->proc, itemvalue);

			if (printed)
				appendStringInfo(&buf, "%s%s", fldsep, value);
			else
				appendStringInfoString(&buf, value);
			printed = true;

			p = att_addlength_pointer(p, typlen, p);
			p = (char *) att_align_nominal(p, typalign);
		}

		/* advance bitmap pointer if any */
		if (bitmap)
		{
			bitmask <<= 1;
			if (bitmask == 0x100)
			{
				bitmap++;
				bitmask = 1;
			}
		}
	}

	PG_RETURN_TEXT_P(PG_STR_GET_TEXT(buf.data));
}

#define HEXBASE 16
/*
 * Convert a int32 to a string containing a base 16 (hex) representation of
 * the number.
 */
Datum
to_hex32(PG_FUNCTION_ARGS)
{
	uint32		value = (uint32) PG_GETARG_INT32(0);
	text	   *result_text;
	char	   *ptr;
	const char *digits = "0123456789abcdef";
	char		buf[32];		/* bigger than needed, but reasonable */

	ptr = buf + sizeof(buf) - 1;
	*ptr = '\0';

	do
	{
		*--ptr = digits[value % HEXBASE];
		value /= HEXBASE;
	} while (ptr > buf && value);

	result_text = PG_STR_GET_TEXT(ptr);
	PG_RETURN_TEXT_P(result_text);
}

/*
 * Convert a int64 to a string containing a base 16 (hex) representation of
 * the number.
 */
Datum
to_hex64(PG_FUNCTION_ARGS)
{
	uint64		value = (uint64) PG_GETARG_INT64(0);
	text	   *result_text;
	char	   *ptr;
	const char *digits = "0123456789abcdef";
	char		buf[32];		/* bigger than needed, but reasonable */

	ptr = buf + sizeof(buf) - 1;
	*ptr = '\0';

	do
	{
		*--ptr = digits[value % HEXBASE];
		value /= HEXBASE;
	} while (ptr > buf && value);

	result_text = PG_STR_GET_TEXT(ptr);
	PG_RETURN_TEXT_P(result_text);
}

/*
 * Create an md5 hash of a text string and return it as hex
 *
 * md5 produces a 16 byte (128 bit) hash; double it for hex
 */
#define MD5_HASH_LEN  32

Datum
md5_text(PG_FUNCTION_ARGS)
{
	text	   *in_text = PG_GETARG_TEXT_PP(0);
	size_t		len;
	char		hexsum[MD5_HASH_LEN + 1];
	text	   *result_text;

	/* Calculate the length of the buffer using varlena metadata */
	len = VARSIZE_ANY_EXHDR(in_text);

	/* get the hash result */
	if (pg_md5_hash(VARDATA_ANY(in_text), len, hexsum) == false)
		ereport(ERROR,
				(errcode(ERRCODE_OUT_OF_MEMORY),
				 errmsg("out of memory")));

	/* convert to text and return it */
	result_text = PG_STR_GET_TEXT(hexsum);
	PG_RETURN_TEXT_P(result_text);
}

/*
 * Create an md5 hash of a bytea field and return it as a hex string:
 * 16-byte md5 digest is represented in 32 hex characters.
 */
Datum
md5_bytea(PG_FUNCTION_ARGS)
{
	bytea	   *in = PG_GETARG_BYTEA_PP(0);
	size_t		len;
	char		hexsum[MD5_HASH_LEN + 1];
	text	   *result_text;

	len = VARSIZE_ANY_EXHDR(in);
	if (pg_md5_hash(VARDATA_ANY(in), len, hexsum) == false)
		ereport(ERROR,
				(errcode(ER