copy.c

PostgreSQL 8.1.4的源码 适用于Linux下的开源数据库系统

/*
 * psql - the PostgreSQL interactive terminal
 *
 * Copyright (c) 2000-2005, PostgreSQL Global Development Group
 *
 * $PostgreSQL: pgsql/src/bin/psql/copy.c,v 1.58 2005/10/15 02:49:40 momjian Exp $
 */
#include "postgres_fe.h"
#include "copy.h"

#include <errno.h>
#include <signal.h>
#include <sys/stat.h>
#ifndef WIN32
#include <unistd.h>				/* for isatty */
#else
#include <io.h>					/* I think */
#endif

#include "libpq-fe.h"
#include "pqexpbuffer.h"
#include "pqsignal.h"

#include "settings.h"
#include "common.h"
#include "prompt.h"
#include "stringutils.h"

#if defined(WIN32) && (!defined(__MINGW32__))
#define __S_ISTYPE(mode, mask)	(((mode) & S_IFMT) == (mask))
#define S_ISDIR(mode)	 __S_ISTYPE((mode), S_IFDIR)
#endif

/*
 * parse_slash_copy
 * -- parses \copy command line
 *
 * The documented preferred syntax is:
 *	\copy tablename [(columnlist)] from|to filename
 *		[ with ] [ oids ] [ delimiter [as] char ] [ null [as] string ]
 * (binary is not here yet)
 *
 * The pre-7.3 syntax was:
 *	\copy tablename [(columnlist)] [with oids] from|to filename
 *		[ [using] delimiters char ] [ with null as string ]
 *
 * The actual accepted syntax is a rather unholy combination of these,
 * plus some undocumented flexibility (for instance, the clauses after
 * WITH can appear in any order).  The accepted syntax matches what
 * the backend grammar actually accepts (see backend/parser/gram.y).
 *
 * table name can be double-quoted and can have a schema part.
 * column names can be double-quoted.
 * filename, char, and string can be single-quoted like SQL literals.
 *
 * returns a malloc'ed structure with the options, or NULL on parsing error
 */

struct copy_options
{
	char	   *table;
	char	   *column_list;
	char	   *file;			/* NULL = stdin/stdout */
	bool		psql_inout;		/* true = use psql stdin/stdout */
	bool		from;
	bool		binary;
	bool		oids;
	bool		csv_mode;
	bool		header;
	char	   *delim;
	char	   *null;
	char	   *quote;
	char	   *escape;
	char	   *force_quote_list;
	char	   *force_notnull_list;
};


static void
free_copy_options(struct copy_options * ptr)
{
	if (!ptr)
		return;
	free(ptr->table);
	free(ptr->column_list);
	free(ptr->file);
	free(ptr->delim);
	free(ptr->null);
	free(ptr->quote);
	free(ptr->escape);
	free(ptr->force_quote_list);
	free(ptr->force_notnull_list);
	free(ptr);
}


/* concatenate "more" onto "var", freeing the original value of *var */
static void
xstrcat(char **var, const char *more)
{
	char	   *newvar;

	newvar = pg_malloc(strlen(*var) + strlen(more) + 1);
	strcpy(newvar, *var);
	strcat(newvar, more);
	free(*var);
	*var = newvar;
}


static struct copy_options *
parse_slash_copy(const char *args)
{
	struct copy_options *result;
	char	   *line;
	char	   *token;
	const char *whitespace = " \t\n\r";

	if (args)
		line = pg_strdup(args);
	else
	{
		psql_error("\\copy: arguments required\n");
		return NULL;
	}

	result = pg_calloc(1, sizeof(struct copy_options));

	token = strtokx(line, whitespace, ".,()", "\"",
					0, false, pset.encoding);
	if (!token)
		goto error;

#ifdef NOT_USED
	/* this is not implemented yet */
	if (pg_strcasecmp(token, "binary") == 0)
	{
		result->binary = true;
		token = strtokx(NULL, whitespace, ".,()", "\"",
						0, false, pset.encoding);
		if (!token)
			goto error;
	}
#endif

	result->table = pg_strdup(token);

	token = strtokx(NULL, whitespace, ".,()", "\"",
					0, false, pset.encoding);
	if (!token)
		goto error;

	/*
	 * strtokx() will not have returned a multi-character token starting with
	 * '.', so we don't need strcmp() here.  Likewise for '(', etc, below.
	 */
	if (token[0] == '.')
	{
		/* handle schema . table */
		xstrcat(&result->table, token);
		token = strtokx(NULL, whitespace, ".,()", "\"",
						0, false, pset.encoding);
		if (!token)
			goto error;
		xstrcat(&result->table, token);
		token = strtokx(NULL, whitespace, ".,()", "\"",
						0, false, pset.encoding);
		if (!token)
			goto error;
	}

	if (token[0] == '(')
	{
		/* handle parenthesized column list */
		result->column_list = pg_strdup(token);
		for (;;)
		{
			token = strtokx(NULL, whitespace, ".,()", "\"",
							0, false, pset.encoding);
			if (!token || strchr(".,()", token[0]))
				goto error;
			xstrcat(&result->column_list, token);
			token = strtokx(NULL, whitespace, ".,()", "\"",
							0, false, pset.encoding);
			if (!token)
				goto error;
			xstrcat(&result->column_list, token);
			if (token[0] == ')')
				break;
			if (token[0] != ',')
				goto error;
		}
		token = strtokx(NULL, whitespace, ".,()", "\"",
						0, false, pset.encoding);
		if (!token)
			goto error;
	}

	/*
	 * Allows old COPY syntax for backward compatibility 2002-06-19
	 */
	if (pg_strcasecmp(token, "with") == 0)
	{
		token = strtokx(NULL, whitespace, NULL, NULL,
						0, false, pset.encoding);
		if (!token || pg_strcasecmp(token, "oids") != 0)
			goto error;
		result->oids = true;

		token = strtokx(NULL, whitespace, NULL, NULL,
						0, false, pset.encoding);
		if (!token)
			goto error;
	}

	if (pg_strcasecmp(token, "from") == 0)
		result->from = true;
	else if (pg_strcasecmp(token, "to") == 0)
		result->from = false;
	else
		goto error;

	token = strtokx(NULL, whitespace, NULL, "'",
					'\\', true, pset.encoding);
	if (!token)
		goto error;

	if (pg_strcasecmp(token, "stdin") == 0 ||
		pg_strcasecmp(token, "stdout") == 0)
	{
		result->psql_inout = false;
		result->file = NULL;
	}
	else if (pg_strcasecmp(token, "pstdin") == 0 ||
			 pg_strcasecmp(token, "pstdout") == 0)
	{
		result->psql_inout = true;
		result->file = NULL;
	}
	else
	{
		result->psql_inout = false;
		result->file = pg_strdup(token);
		expand_tilde(&result->file);
	}

	token = strtokx(NULL, whitespace, NULL, NULL,
					0, false, pset.encoding);

	/*
	 * Allows old COPY syntax for backward compatibility 2002-06-19
	 */
	if (token && pg_strcasecmp(token, "using") == 0)
	{
		token = strtokx(NULL, whitespace, NULL, NULL,
						0, false, pset.encoding);
		if (!(token && pg_strcasecmp(token, "delimiters") == 0))
			goto error;
	}
	if (token && pg_strcasecmp(token, "delimiters") == 0)
	{
		token = strtokx(NULL, whitespace, NULL, "'",
						'\\', false, pset.encoding);
		if (!token)
			goto error;
		result->delim = pg_strdup(token);
		token = strtokx(NULL, whitespace, NULL, NULL,
						0, false, pset.encoding);
	}

	if (token)
	{
		/*
		 * WITH is optional.  Also, the backend will allow WITH followed by
		 * nothing, so we do too.
		 */
		if (pg_strcasecmp(token, "with") == 0)
			token = strtokx(NULL, whitespace, NULL, NULL,
							0, false, pset.encoding);

		while (token)
		{
			bool		fetch_next;

			fetch_next = true;

			/* someday allow BINARY here */
			if (pg_strcasecmp(token, "oids") == 0)
				result->oids = true;
			else if (pg_strcasecmp(token, "csv") == 0)
				result->csv_mode = true;
			else if (pg_strcasecmp(token, "header") == 0)
				result->header = true;
			else if (pg_strcasecmp(token, "delimiter") == 0)
			{
				token = strtokx(NULL, whitespace, NULL, "'",
								'\\', false, pset.encoding);
				if (token && pg_strcasecmp(token, "as") == 0)
					token = strtokx(NULL, whitespace, NULL, "'",
									'\\', false, pset.encoding);
				if (token)
					result->delim = pg_strdup(token);
				else
					goto error;
			}
			else if (pg_strcasecmp(token, "null") == 0)
			{
				token = strtokx(NULL, whitespace, NULL, "'",
								'\\', false, pset.encoding);
				if (token && pg_strcasecmp(token, "as") == 0)
					token = strtokx(NULL, whitespace, NULL, "'",
									'\\', false, pset.encoding);
				if (token)
					result->null = pg_strdup(token);
				else
					goto error;
			}
			else if (pg_strcasecmp(token, "quote") == 0)
			{
				token = strtokx(NULL, whitespace, NULL, "'",
								'\\', false, pset.encoding);
				if (token && pg_strcasecmp(token, "as") == 0)
					token = strtokx(NULL, whitespace, NULL, "'",
									'\\', false, pset.encoding);
				if (token)
					result->quote = pg_strdup(token);
				else
					goto error;
			}
			else if (pg_strcasecmp(token, "escape") == 0)
			{
				token = strtokx(NULL, whitespace, NULL, "'",
								'\\', false, pset.encoding);
				if (token && pg_strcasecmp(token, "as") == 0)
					token = strtokx(NULL, whitespace, NULL, "'",
									'\\', false, pset.encoding);
				if (token)
					result->escape = pg_strdup(token);
				else
					goto error;
			}
			else if (pg_strcasecmp(token, "force") == 0)
			{
				token = strtokx(NULL, whitespace, ",", "\"",
								0, false, pset.encoding);
				if (pg_strcasecmp(token, "quote") == 0)
				{
					/* handle column list */
					fetch_next = false;
					for (;;)
					{
						token = strtokx(NULL, whitespace, ",", "\"",
										0, false, pset.encoding);
						if (!token || strchr(",", token[0]))
							goto error;
						if (!result->force_quote_list)
							result->force_quote_list = pg_strdup(token);
						else
							xstrcat(&result->force_quote_list, token);
						token = strtokx(NULL, whitespace, ",", "\"",
										0, false, pset.encoding);
						if (!token || token[0] != ',')
							break;
						xstrcat(&result->force_quote_list, token);
					}
				}
				else if (pg_strcasecmp(token, "not") == 0)
				{
					token = strtokx(NULL, whitespace, ",", "\"",
									0, false, pset.encoding);
					if (pg_strcasecmp(token, "null") != 0)
						goto error;
					/* handle column list */
					fetch_next = false;