This is fairly confusing, but I'll try to explain as best I can...
I've got a MYSQL table full of strings like this:
{3}12{2}3{5}52
{3}7{2}44
{3}15{2}2{4}132{5}52{6}22
{3}15{2}3{4}168{5}52
Each string is a combination of product options and option values. The numbers inside the { } are the option, for example {3} = Color. The number immediately following each { } number is that option's value, for example 12 = Blue. I've already got the PHP code that knows how to parse these strings and deliver the information correctly, with one exception: For reasons that are probably too convoluted to get into here, the order of the options needs to be 3,4,2,5,6. (To try to modify the rest of the system to accept the current order would be too monumental a task.) It's fine if a particular combination doesn't have all five options, for instance "{3}7{2}44" delivers the expected result. The problem is just with combinations that include option 2 AND option 4-- their order needs to be switched so that any combination that includes both options 2 and 4, the {4} and its corresponding value comes before the {2} and it's corresponding value.
I've tried bringing the column into Excel and using Text to Columns, splitting them up by the "{" and "}" characters and re-ordering the columns, but since not every string yields the same number of columns, the order gets messed up in other ways (like option 5 coming before option 2).
I've also experimented with using PHP to explode each string into an array (which I thought I could then re-sort) using "}" as the delimiter, but I had no luck with that either because then the numbers blend together in other ways that make them unusable.
TL;DR: I have a bunch of strings like the ones quoted above. In every string that contains both a "{2}" and a "{4}", the placement of both of those values needs to be switched, so that the {4} and the number that follows it comes before the {2} and the number that follows it. In other words:
{3}15{2}3{4}168{5}52
needs to become
{3}15{4}168{2}3{5}52
The closest I've been able to come to a solution, in pseudocode, would be something like:
for each string,
if "{4}" is present in this string AND "{2}" is present in this string,
take the "{4}" and every digit that follows it UNTIL you hit another "{" and store that substring as a variable, then remove it from the string.
then, insert that substring back into the string, at a position starting immediately before the "{2}".
I hope that makes some kind of sense...
Is there any way with PHP, Excel, Notepad++, regular expressions, etc., that I can do this? Any help would be insanely appreciated.
EDITED TO ADD: After several people posted solutions, which I tried, I realized that it would be crucial to mention that my host is running PHP 5.2.17, which doesn't seem to allow for usort with custom sorting. If I could upvote everyone's solution (all of which I tried in PHP Sandbox and all of which worked), I would, but my rep is too low.
How would something like this work for you. The first 9 lines just transform your string into an array with each element being an array of the option number and value. The Order establishes an order for the items to appear in and the last does a usort utilizing the order array for positions.
$str = "{3}15{2}2{4}132{5}52{6}22";
$matches = array();
preg_match_all('/\{([0-9]+)\}([0-9]+)/', $str, $matches);
array_shift($matches);
$options = array();
for($x = 0; $x < count($matches[0]); $x++){
$options[] = array($matches[0][$x], $matches[1][$x]);
}
$order = [3,4,2,5,6];
usort($options, function($a, $b) use ($order) {
return array_search($a[0], $order) - array_search($b[0], $order);
});
To get you data back into the required format you would just
$str = "";
foreach($options as $opt){
$str.="{".$opt[0]."}".$opt[1];
}
On of the bonuses here is that when you add a new options type inserting adjusting the order is just a matter of inserting the option number in the correct position of the $order array.
First of all, those options should probably be in a separate table. You're breaking all kinds of normalization rules stuffing those things into a string like that.
But if you really want to parse that out in php, split the string into a key=>value array with something like this:
$options = [];
$pairs = explode('{', $option_string);
foreach($pairs as $pair) {
list($key,$value) = explode('}', $pair);
$options[$key] = $value;
}
I think this will give you:
$options[3]=15;
$options[2]=3;
$options[4]=168;
$options[5]=52;
Another option would be to use some sort of existing serialization (either serialize() or json_encode() in php) instead of rolling your own:
$options_string = json_encode($options);
// store $options_string in db
then
// get $options_string from db
$options = json_decode($options_string);
Here's a neat solution:
$order = array(3, 4, 2, 5, 6);
$string = '{3}15{2}3{4}168{5}52';
$split = preg_split('#\b(?={)#', $string);
usort($split, function($a, $b) use ($order) {
$a = array_search(preg_replace('#^{(\d+)}\d+$#', '$1', $a), $order);
$b = array_search(preg_replace('#^{(\d+)}\d+$#', '$1', $b), $order);
return $a - $b;
});
$split = implode('', $split);
var_dump($split);
Related
The following is the code
<?php
$id ="202883-202882-202884-0";
$str = implode('-',array_unique(explode('-', $id)));
echo $str;
?>
The result is
202883-202882-202884-0
for $id ="202883-202882-202882-0";, result is 202883-202882-0
I would like to replace the duplicate value with zero, so that the result should be like 202883-202882-0-0, not just remove it.
and for $id ="202883-0-0-0";, result should be 202883-0-0-0. zero should not be replaced, repeating zeros are allowed.
How can I archive that?
More info:
I want to replace every duplicate numbers. Because this is for a product comparison website. There will be only maximum 4 numbers. each will be either a 6 digit number or single digit zero. all zero means no product was selected. one 6 digit number and 3 zero means, one product selected and 3 blank.
Each 6 digit number will collect data from database, I dont want to allow users to enter same number multiple times (will happen only if the number is add with the URL manually.).
Update: I understand that my question was not clear, may be my English is poor.
Here is more explanation, this function is for a smartphone comparison website.
The URL format is sitename.com/compare.html?id=202883-202882-202889-202888.
All three numbers are different smartphones(their database product ID).
I dont want to let users to type in the same product ID like id=202883-202882-202882-202888. It will not display two 202882 results in the website, but it will cause some small issues. The URL will be same without change, but the internal PHP code should consider it as id=202883-202882-202888-0.
The duplicates should be replaced as zero and added to the end.
There will be only 4 numbers separated by "-".
The following examples might clear the cloud!
if pid=202883-202882-202889-202888 the result should be 202883-202882-202889-202888
if pid=202883-202883-202883-202888 the result should be 202888-0-0-0
if pid=202883-202882-202883-202888 the result should be 202883-202882-202888-0
if pid=202882-202882-202882-202882 the result should be 202882-0-0-0
I want to allow only either 6 digit numbers or single digit zero through the string.
if pid=rgfsdg-fgsdfr4354-202883-0 the result should be 202883-0-0-0
if pid=fasdfasd-asdfads-adsfds-dasfad the result should be 0-0-0-0
if pid=4354-45882-445202882-202882 the result should be 202882-0-0-0
It is too complicated for me create, I know there are bright minds out there who can do it much more efficiently than I can.
You can do a array_unique (preserves key), then fill the gaps with 0. Sort by key and you are done :)
+ on arrays will unify the arrays but prioritizes the one on the left.
Code
$input = "0-1-1-3-1-1-3-5-0";
$array = explode('-', $input);
$result = array_unique($array) + array_fill(0, count($array), 0);
ksort($result);
var_dump(implode('-',$result));
Code (v2 - suggested by mickmackusa) - shorter and easier to understand
Fill an array of the size of the input array. And replace by leftover values from array_unique. No ksort needed. 0s will be replaced at the preserved keys of array_unique.
$input = "0-1-1-3-1-1-3-5-0";
$array = explode('-', $input);
$result = array_replace(array_fill(0, count($array), 0), array_unique($array));
var_export($result);
Working example.
Output
string(17) "0-1-0-3-0-0-0-5-0"
Working example.
references
ksort - sort by key
array_fill - generate an array filled with 0 of a certain length
This is another way to do it.
$id = "202883-202882-202882-0-234567-2-2-45435";
From the String you explode the string into an array based on the delimiter which in this case is '-'/
$id_array = explode('-', $id);
Then we can loop through the array and for every unique entry we find, we can store it in another array. Thus we are building an array as we search through the array.
$id_array_temp = [];
// Loop through the array
foreach ($id_array as $value) {
if ( in_array($value, $id_array_temp)) {
// If the entry exists, replace it with a 0
$id_array_temp[] = 0;
} else {
// If the entry does not exist, save the value so we can inspect it on the next loop.
$id_array_temp[] = $value;
}
}
At the end of this operation we will have an array of unique values with any duplicates replaced with a 0.
To recreate the string, we can use implode...
$str = implode('-', $id_array_temp);
echo $str;
Refactoring this, using a ternary to replace the If,else...
$id_array = explode('-', $id);
$id_array_temp = [];
foreach ($id_array as $value) {
$id_array_temp[] = in_array($value, $id_array_temp) ? 0 : $value;
}
$str = implode('-', $id_array_temp);
echo $str;
Output is
202883-202882-0-0-234567-2-0-45435
This appears to be a classic XY Problem.
The essential actions only need to be:
Separate the substrings in the hyphen delimited string.
Validate that the characters in each substring are in the correct format AND are unique to the set.
Only take meaningful action on qualifying value.
You see, there is no benefit to replacing/sanitizing anything when you only really need to validate the input data. Adding zeros to your input just creates more work later.
In short, you should use a direct approach similar to this flow:
if (!empty($_GET['id'])) {
$ids = array_unique(explode('-', $_GET['id']));
foreach ($ids as $id) {
if (ctype_digit($id) && strlen($id) === 6) {
// or: if (preg_match('~^\d{6}$~', $id)) {
takeYourNecessaryAction($id);
}
}
}
I'm not sure the title really gets across what I'm asking, so here's what I'm trying to do:
I have an array of arrays with four integer elements each, ie.
Array(Array(1,2,3,4), Array(4,2,3,1), Array(18,3,22,9), Array(23, 12, 33, 55))
I basically need to remove one of the two arrays that have the same values in any order, like indices 0 and 1 in the example above.
I can do this pretty easily when there are only two elements to check, using the best answer code in this question.
My multidimensional array can have 1-10 arrays at any given time, so I can't seem to figure out the best way to process a structure like that and remove arrays that have the same numbers in any order.
Thanks so much!
I've been thinking about this, and I think using a well designed closure with array_filter might be the way I'd go about this:
$matches = array();
$array = array_filter($array, function($ar) use (&$matches) {
sort($ar);
if(in_array($ar, $matches)) {
return false;
}
$matches[] = $ar;
return true;
});
See here for an example: http://ideone.com/Zl7tlR
Edit: $array will be your final result, ignore $matches as it's just used during the filter closure.
I'm build a site that is not in English, but some of the data may contain English. When I sort lists, for example:
sort($tags, SORT_LOCALE_STRING);
English strings always come first (because Latin letters have the lowest ordinals in Unicode), even though my site is primarily for non-English speakers.
I would like to have a sort function that sorts the strings lexicographically in the same way as the standard PHP sort, but always puts English strings after all the non-English strings. Any ideas?
Ok, this is cheating, but eventually I did this:
$tags = getSomeTags();
$tags_eng = $tags_lang = [];
foreach ($tags as $tag => $value) {
if ($tag[0] <= 'z') $tags_eng[$tag] = $value;
else $tags_lang[$tag] = $value;
}
ksort($tags_eng); ksort($tags_lang);
$tags = $tags_lang + $tags_eng;
Array "union" operation maintains the order of elements, so that works well.
Please note that its important not to use array_merge() for concatenating the arrays at the end, because if you have string keys that look like a number (compare "9" with 9), then array_merge() will think its an "indexed item" and will do weird stuff with it (resetting the key to the earliest available index).
I'd love it if someone can come up with a better method.
Suppose a string:
$str = 'a_b_c';
I want match all possible combination with a, b, c with above. For example:
b_a_c, c_a_b, a_c_b..etc will be give true when compare with above $str.
NOTE:
$str may be random. eg: a_b, k_l_m_n etc
I would split your string into an array, and then compare it to an array of elements to match on.
$originalList = explode('_', 'a_b_c');
$matchList = array('a', 'b', 'c');
$diff = array_diff($matchList, $originalList);
if (!empty($diff)) {
// At least one of the elements in $matchList is not in $originalList
}
Beware of duplicate elements and what not, depending on how your data comes in.
Documentation:
array_diff()
explode()
There is no builtin way to quickly do this. Your task can be accomplished many different ways which will vary on how general they are. You make no mention of null values or checking the formatting of the string, so something like this might work for your purpose:
function all_combos($str,$vals) {
$s=explode("_",$str);
foreach($s as $c) {
if(!in_array($s,$vals)) return false;
}
return true;
}
Call like all_combos("b_c_a",array("a","b","c"));
I have a comma delimited string held within a database field that could contain any number of values:
23,45,21,40,67,22
I need to be able to somehow switch two values, so for example I know I need to move 45 one position down the string, so I end up with:
23,21,45,40,67,22
The reason for this is that the numbers all correspond to the IDs held in another database table, and their position in the sting determine the order those items will be printed on screen. Before you ask about database design - I've inherited it and it cannot be changed without significant work to an entire application.
So I've thought about exploding the string, identifying the position of the target number and swapping it with the one next-door, but I'm unsure of how this can be achieved when the total number of values is not known.
Any things? I suspect the solution will be cumbersome, but needs must!!
assuming you need to only move the desired value down one position in the array:
$values = explode(',', $data_string);
$value_to_move = 45;
$value_count = count($values);
for($i=0;$i<$value_count;$i++)
{
if($values[$i] == $value_to_move)
{
if($i < ($value_count-1))
{ // if the value to move is NOT at the end of the list already
$values[$i] = $values[$i+1];
$values[$i+1] = $value_to_move;
$i++;
}
}
}
$new_data_string = implode(',', $values);
I'd just pull them into an array and work with them there. Write the string out in comma-delimited format again, and rewrite that to the DB.
Assuming you know exactly which two values to switch in that list, then explode is the best option:
$array = explode(',', $string)
# find the two values (NOTE: *NO* error handling, what if the values aren't there?)
$index1 = array_search($first_value, $array);
$index2 = array_search($second_value, $array);
# swap them
$temp = $array[$index1];
$array[$index1] = $array[$index2];
$array[$index2] = $temp;
# rebuild the array
$string = implode(',', $array);