[Ernest Marcinko]

Hi,

Thank you for the details!

I think the issue might be the auto-populate setting, you seem to have set it to display the latest results, which overrides the primary ordering for the initial results.

Can you please try this setup instead: https://i.imgur.com/u32EyJS.png

It will do the trick.

[Ernest Marcinko]

Sure, let me know 🙂 It should do the trick, I see no other issues there.

[Ernest Marcinko]

I forgot to add, I think if you simply open and save the old publication without changing them, the values get automatically converted by ACF. So if there is only a few old publications, that’s the easiest and fastest fix.

[Ernest Marcinko]

Okay, I have found the issue, it is however not related to the search.

The filters were working fine, which I found strange, so I started checking the actual values of the publish_date field and found the culprit.

The issue is that some values of that field are stored as timestamps and some values are stored as ACF dates. If you take a look at these debug outputs, in this one the ACF date values are stored, but on other entries (I guess older entries) the timestamp is stored.

I suspect this field was imported and the data types are not matching with the current ACF date type. The only way to resolve this is to convert the incorrect timestamp data to ACF date data type. (especially if you want to use ACF date field) For display purposes it’s fine, ACF and Ajax Search Pro can both convert them, but for searching it is a massive issue, as there is no way to query on both at the same time, the data types must be the same.

[Ernest Marcinko]

Thank you, it works. I will do a bit of debugging now, don’t get scared if you see some weird output 🙂

[Ernest Marcinko]

Thanks!

I have turned off the predictive keyword suggestions, now it returns no results as expected. Don’t forget to exclude the sizes.

[Ernest Marcinko]

Hi,

Thank you very much for the details, it helps a lot!

It seems like there might be some sort of an unknown bug – when both filters are selected, it works. For now I have set relative values to both filters to make it work, but I would like to debug this issue more. I suspect it has to do something with the empty value on the second filter, somehow overriding the first due to the same field name.

I tried to log-in via (s)FTP, but it does not authenticate, can you please check that?

[Ernest Marcinko]

Hi Bill!

If the live loader for woocommerce results page is enabled, the plugin will try to detect if the current page is the shop page, and if so, then will try to load the results through the shop page – like it’s in the tutorial.

However, if you place the same search shortcode to a non-shop page, then it will use it’s own results container. So there is a detection method, and if the shop archive is not found it falls back to it’s results container.

If you are using a page builder of some sort that has a pop-up feature, you can place the search into that, so more-less you can achieve something similar.
I have recently made a video tutorial on a woocommerce specific search example, I definitely recommend checking it here, I think you will find some good information in it.

[Ernest Marcinko]

Hi,

It’s because of the amount of text you are printing to the page and probably in combination with the keyword highlighter. I’m looking at the extracted texts, some of those are really really long and it eats up browser memory. It isn’t related to the search process, but the amount of data pushed to the page, it is way too much.

I would suggest turning off the keyword highlighter first, to see if that changes anything. It would explain the lag, as it has to parse through thousands of pages on page load and the browser memory spikes through the roof. If that doesn’t help at all, then it’s a script somewhere in elementor that tries to deal with the data and just fails. Then try to completely remove the extracted text, that should immediately resolve the problem.

The best solution would be to not load the text during page load, but to mitigate it to an asynchronous handler via rest API to load the text when the user clicks on the “view extracted text” icon. Nothing too complicated, only a single rest endpoint and a handler, that should do the trick. Not sure if Elementor has support for loading data asynchronously though, you may have to custom code it.

[Ernest Marcinko]

You cannot access this content.

[Ernest Marcinko]

Do you have an issue with the plugin feature that should work and it doesn’t? That policy covers security issues or major problems with plugin features that can’t be fixed.

Can you please describe the problem in details, add temporary log-in details so we can check and rectify the problem? Thank you!

• This reply was modified 1 year, 1 month ago by Ernest Marcinko.

[Ernest Marcinko]

You cannot access this content.

[Ernest Marcinko]

You cannot access this content.

[Ernest Marcinko]

Hi,

Thank you very much for the details, it helps a lot!

From looking at the front-end, I suspect the minimum character trigger is set to 3 or more characters instead of the default 0, can you please make sure to set it to 0 here: https://i.imgur.com/RT2xrgb.png
That will resolve the issue.

As for the publications, while there is no option for that, it should be still possible with a small custom snippet:

add_filter(
	'asp_results',
	function ( $results ) {
		foreach ( $results as $r ) {
			if ( $r->content_type !== 'attachment' ) {
				continue;
			}
			$parent_id = wp_get_post_parent_id( $r->id );
			if ( !is_wp_error($parent_id) ) {
				$r->title = get_the_title($parent_id);
			}
		}
		return $results;
	},
	10,
	1
);

Try adding this code via the Code Snippets plugin or to the functions.php file in your theme/child theme directory – make sure to have a full server back-up first for safety. For more details you can check the safe coding guidelines.

[Ernest Marcinko]

You cannot access this content.

[Author]

Posts