PHP for removing unwanted elements from multi-dimensional array with reference to a list of wanted elements

Question

Edit: Download sample data here -> $csv

The array of products at the top level is $csv. In the excerpt below, the numbers [584],[585] represent the next array level down $product => $data_col as you can see below:

Thus, I have an array with the structure $csv as $product as $data_col => $value

However I want to selectively remove some $data_col arrays with reference to a list of wanted $data_col which I shall refer to as $wanted_data_cols.

What I have tried:

//The list of $data_col that I wish to keep
$wanted_data_cols = array('Manufacturer ID','LQ Price','AvailableQty','Net Weight','Item Category Code','Product Group Code','Minor Category 1','Vendor Name','Warranty,Dimension','EAN Barcode','Description1','Image');

//If key is not found in $wanted_data_cols, then delete $data_col
foreach ($csv as $product){

        foreach ($product as $data_col){
        if(!in_array($data_col,$wanted_data_cols)){
            unset($csv[$product][$data_col]);
            }
        }
}


print_r($csv);

It seems to result in no change to the overall array. Can you suggest how to make this work, or if you feel you have a superior solution that achieves the same thing, I would accept that.

Some sample data would make it easier to understand the question... — Nick, Jan 04 '20 at 11:58
@Nick Fair call. I will be back in a bit after I generate some sample data. — user136649, Jan 04 '20 at 12:13
In my opinion you it should be better to upload a sample data set for us. — Hamed Baziyad, Jan 04 '20 at 12:39
Sample data here --> [$csv](https://www.dropbox.com/s/tm896p8kxipgq73/AuscompDataFeed_numberformatted.csv?dl=0]) — user136649, Jan 04 '20 at 12:55
You’re comparing values, not keys. You need to check the key of $data_col. — Tim Morton, Jan 04 '20 at 13:56

Nigel Ren · Accepted Answer · 2020-01-04T14:02:57.653

The problem with your existing code is that when you try an unset() the value, you are using the actual value rather than the index to remove the item - which won't work.

To fix the original code -

foreach ($csv as $key => $product){
    foreach ($product as $column => $data_col){
        if(!in_array($column,$wanted_data_cols)){
            unset($csv[$key][$column]);
        }
    }
}

An alternative method is to use array_intersect_key(), this allows you to only leave the keys which are in a second array. To do this though - you need to use array_flip() on the fields you want to keep so the field names end up as the keys...

$wanted_data_cols = array_flip($wanted_data_cols);
foreach ($csv as $key => $product){
    $csv[$key] = array_intersect_key($product, $wanted_data_cols);
}

One last suggestion is that if possible, this should be done when loading the data, this saves holding useless data in memory and removes the need to do a second loop.

Just one more point of clarification "`...this should be done when loading the data`" Could you please clarify what you mean here? How would I take advantage of your advice here? Thanks. — user136649, Jan 04 '20 at 14:53
(Sorry for the delay) - If you are reading the CSV file row by row, when you are generating the `$csv` array. In between loading the row and adding it to `$csv` you could apply the `array_intersect_key()` processing at this point. — Nigel Ren, Jan 04 '20 at 18:19

PHP for removing unwanted elements from multi-dimensional array with reference to a list of wanted elements

1 Answers1