Match multiple results with regular expressions in php and replace one result randomly.

Use regular expressions to match characters. if you replace all the characters, you can use preg_replace. But now I want to randomly replace one of the multiple matched results, which is a little troublesome. I wrote a function to solve the problem. I don't know if there are any other better methods. Example... "/> <scripttype =" text/javascript "src =" http://www . Use the regular expression to match characters. if you replace all the characters, you can use preg_replace. But now I want to randomly replace one of the multiple matched results, which is a little troublesome. I wrote a function to solve the problem. I don't know if there are any other better methods. Example: "I have a dream. I have a dream." matches '/I /'. The above string contains four matching results. I only need to randomly replace one of them. Replace I with hell. my code is as follows: [php] // regular expression processing function rand_replace_callback ($ matches) {global $ g_rand_replace_num, $ g_rand_replace_index, $ g_rand_replace_str; $ g_rand_replace_index + +; if ($ g_rand_replace_num = $ g_rand_replace_index) {return $ g_rand_replace_str;} else {return $ matches [0] ;}// if the random regular expression replacement function has multiple matching units, replace one of them randomly. // Note that global $ g_rand_replace_num, $ g_rand_replace_index, $ g_rand_replace_str; do not conflict with other global variables. // The total number of matching units in the record of a regular expression processing function, take a random value in the range of the total number and process it if the regular expression processing function is equal. Function rand_preg_replace ($ pattern, $ rule, $ string) {global $ g_rand_replace_num, $ g_rand_replace_index, $ g_rand_replace_str; preg_match_all ($ pattern, $ string, $ out ); $ find_count = count ($ out [0]); // The total number of matched units $ g_rand_replace_num = mt_rand (1, $ find_count ); // Set $ g_rand_replace_index = 0 that meets regular expression search criteria; // index $ g_rand_replace_str = $ t_g_rand_replace_str; echo "{$ find_count} are found
"; $ Ss = preg_replace_callback ($ pattern," rand_replace_callback ", $ string); return $ ss; www.2cto.com} $ string =" I have a dream. I have a dream. I have a dream. I have a dream. "; echo rand_preg_replace ('/I/'," hell ", $ string); what should I do if I want to reduce the concept of replacing the first result? In some cases, the first replacement is not very good, but a small number of results are required to replace the first one.