搜索
首页php教程php手册解析strtr函数的效率问题

解析strtr函数的效率问题

Jun 13, 2016 am 11:44 AM
strstrtr一般使用函数匹配字符串操作效率替换解析进行问题

最近经常要对字符串进行匹配和替换操作,之前一般使用str_replace或者preg_replace,据说strtr的效率不错,所以对比了一下:

复制代码 代码如下:


$i = 0;
$t = microtime(true);
for(;$i{
    $str = strtr(md5($i), $p2);
}
var_dump(microtime(true)-$t);    //0.085476875305176
$t = microtime(true);
for(;$i{
    $str = preg_replace($p, '', md5($i));
}
var_dump(microtime(true)-$t);   //0.09863805770874


结果显示,strtr的效率比preg_replace高约15%左右。
趁着周末,查看了strtr的php源码:

复制代码 代码如下:


PHP_FUNCTION(strtr)
{
        zval **str, **from, **to;
        int ac = ZEND_NUM_ARGS();
        //参数检查(zend_get_parameters_ex函数定义在zend_api.c文件中)
        if (ac 3 || zend_get_parameters_ex(ac, &str, &from, &to) == FAILURE) {
                WRONG_PARAM_COUNT;
        }
        //参数检查
        if (ac == 2 && Z_TYPE_PP(from) != IS_ARRAY) {
                php_error_docref(NULL TSRMLS_CC, E_WARNING, "The second argument is not an array.");
                RETURN_FALSE;
        }
        convert_to_string_ex(str);
        /* shortcut for empty string */
        //宏Z_STRLEN_PP定义在zend_operators.h
        if (Z_STRLEN_PP(str) == 0) {
                RETURN_EMPTY_STRING();
        }
        if (ac == 2) {
                php_strtr_array(return_value, Z_STRVAL_PP(str), Z_STRLEN_PP(str), HASH_OF(*from));
        } else {
                convert_to_string_ex(from);
                convert_to_string_ex(to);
                ZVAL_STRINGL(return_value, Z_STRVAL_PP(str), Z_STRLEN_PP(str), 1);
                php_strtr(Z_STRVAL_P(return_value),
                                  Z_STRLEN_P(return_value),
                                  Z_STRVAL_PP(from),
                                  Z_STRVAL_PP(to),
                                  MIN(Z_STRLEN_PP(from),
                                  Z_STRLEN_PP(to)));
        }
}


先看看php_strtr函数:

复制代码 代码如下:


//trlen是字符串str_from与str_to的长度的最小值
PHPAPI char *php_strtr(char *str, int len, char *str_from, char *str_to, int trlen)
{
        int i;
        unsigned char xlat[256]; //
        if ((trlen                 return str;
        }
        //xlat的下标与值相等
        for (i = 0; i         //把from到to字符串的每一个字符对应起来。例如:from="ab",to="cd",则会产生这样的对应'a'=>'c', 'b'=>'d'。
        for (i = 0; i                 xlat[(unsigned char) str_from[i]] = str_to[i];
        }
        //替换(不过觉得这个函数的效率还有可以改进的地方,因为如果需要替换的字符只是占整个字符串很少的部分,这样就有大部分的赋值操作其实并没有什么意义,这样的情况下感觉先判断再赋值感觉会高效一点。有空测试一下)
        for (i = 0; i                 str[i] = xlat[(unsigned char) str[i]];
        }
        return str;
}


可见,在处理strtr('abcdaaabcd', 'ab', 'efd')这样的操作时,应该是很高效的。
(注意:这个操作输出efcdeeefcd) 
再看看php_strtr_array:

复制代码 代码如下:


static void php_strtr_array(zval *return_value, char *str, int slen, HashTable *hash)
{
        zval **entry;
        char  *string_key;
        uint   string_key_len;
        zval **trans;
        zval   ctmp;
        ulong num_key;
        int minlen = 128*1024;
        int maxlen = 0, pos, len, found;
        char *key;
        HashPosition hpos;
        smart_str result = {0};
        HashTable tmp_hash;
        //把替换数组从hash复制到tmp_hash,并记录下标字符串的最大和最小长度
        zend_hash_init(&tmp_hash, 0, NULL, NULL, 0);
        zend_hash_internal_pointer_reset_ex(hash, &hpos);
        while (zend_hash_get_current_data_ex(hash, (void **)&entry, &hpos) == SUCCESS) {
                switch (zend_hash_get_current_key_ex(hash, &string_key, &string_key_len, &num_key, 0, &hpos)) {
                        case HASH_KEY_IS_STRING:
                                len = string_key_len-1;
                                if (len                                         zend_hash_destroy(&tmp_hash);
                                        RETURN_FALSE;
                                }
                                zend_hash_add(&tmp_hash, string_key, string_key_len, entry, sizeof(zval*), NULL);
                                if (len > maxlen) {
                                        maxlen = len;
                                }
                                if (len                                         minlen = len;
                                }
                                break;
                        //下标如果是整形的话会转换成字符串类型,例如:array(10=>'aa')转换成array('10'=>'aa')
                        case HASH_KEY_IS_LONG:
                                Z_TYPE(ctmp) = IS_LONG;
                                Z_LVAL(ctmp) = num_key;
                                convert_to_string(&ctmp);
                                len = Z_STRLEN(ctmp);
                                zend_hash_add(&tmp_hash, Z_STRVAL(ctmp), len+1, entry, sizeof(zval*), NULL);
                                zval_dtor(&ctmp);
                                if (len > maxlen) {
                                        maxlen = len;
                                }
                                if (len                                         minlen = len;
                                }
                                break;
                }
                zend_hash_move_forward_ex(hash, &hpos);
        }
        key = emalloc(maxlen+1);
        pos = 0;
        //从字符串的第一个字符开始循环匹配,pos记录当前查找的位置
        while (pos                 //当前位置加上最大长度,如果大于字符串长度,则最大长度就需要改变
                if ((pos + maxlen) > slen) {
                        maxlen = slen - pos;
                }
                found = 0;
                memcpy(key, str+pos, maxlen);
                //从最大长度开始匹配,就是说对'abcd',若array('a'=>'e','ab'=>'f'),则会先把ab替换为f,而不是先把a换成e。
                for (len = maxlen; len >= minlen; len--) {
                        key[len] = 0;
                        //因为使用了hash表,所以这样的效率还是挺高的
                        if (zend_hash_find(&tmp_hash, key, len+1, (void**)&trans) == SUCCESS) {
                                char *tval;
                                int tlen;
                                zval tmp;
                                if (Z_TYPE_PP(trans) != IS_STRING) {
                                        tmp = **trans;
                                        zval_copy_ctor(&tmp);
                                        convert_to_string(&tmp);
                                        tval = Z_STRVAL(tmp);
                                        tlen = Z_STRLEN(tmp);
                                } else {
                                        tval = Z_STRVAL_PP(trans);
                                        tlen = Z_STRLEN_PP(trans);
                                }
                                //加入结果
                                smart_str_appendl(&result, tval, tlen);
                                //向前跳跃
                                pos += len;
                                found = 1;
                                if (Z_TYPE_PP(trans) != IS_STRING) {
                                        zval_dtor(&tmp);
                                }
                                break;
                        }
                }
                if (! found) {
                        smart_str_appendc(&result, str[pos++]);
                }
        }
        efree(key);
        zend_hash_destroy(&tmp_hash);
        smart_str_0(&result);
        RETVAL_STRINGL(result.c, result.len, 0);
}


声明
本文内容由网友自发贡献,版权归原作者所有,本站不承担相应法律责任。如您发现有涉嫌抄袭侵权的内容,请联系admin@php.cn

热AI工具

Undresser.AI Undress

Undresser.AI Undress

人工智能驱动的应用程序,用于创建逼真的裸体照片

AI Clothes Remover

AI Clothes Remover

用于从照片中去除衣服的在线人工智能工具。

Undress AI Tool

Undress AI Tool

免费脱衣服图片

Clothoff.io

Clothoff.io

AI脱衣机

Video Face Swap

Video Face Swap

使用我们完全免费的人工智能换脸工具轻松在任何视频中换脸!

热工具

记事本++7.3.1

记事本++7.3.1

好用且免费的代码编辑器

Atom编辑器mac版下载

Atom编辑器mac版下载

最流行的的开源编辑器

MinGW - 适用于 Windows 的极简 GNU

MinGW - 适用于 Windows 的极简 GNU

这个项目正在迁移到osdn.net/projects/mingw的过程中,你可以继续在那里关注我们。MinGW:GNU编译器集合(GCC)的本地Windows移植版本,可自由分发的导入库和用于构建本地Windows应用程序的头文件;包括对MSVC运行时的扩展,以支持C99功能。MinGW的所有软件都可以在64位Windows平台上运行。

禅工作室 13.0.1

禅工作室 13.0.1

功能强大的PHP集成开发环境

WebStorm Mac版

WebStorm Mac版

好用的JavaScript开发工具