Category: Computers & Technology - Page 22

How to sort empty values last in WordPress

For the past several days I’ve been hammering my head against a conundrum: how to get WordPress to sort a set of posts in ascending order, but with empty values at the end of the list instead of the beginning.

This seems like it should be a simple option in the query. But MySQL doesn’t offer a straightforward way to do this. There are some fairly simple MySQL tricks that will accomplish it, but there’s no way to apply those tricks within the context of WP_Query because they require manipulating either the SELECT or ORDER BY portions of the SQL query in ways WP_Query doesn’t allow. (I mean, you can write custom SQL for WP_Query, but if you’re trying to alter the output of the main query, good luck.)

I tried everything I could possibly think of yesterday with the pre_get_posts hook, but it all went nowhere, other than discovering a very weird quirk of MySQL that I don’t fully understand and won’t bother explaining here.

Sleep on it

I woke up this morning with an idea! I resigned myself to the fact that this ordering can’t happen before the query runs, but I should be able to write a pretty simple function to do it after the query has run.

Bear one key thing in mind: This is not going to work properly with paginated results. I mean, it’ll sort of work. The empty values will get sorted to the end of the list, but they’ll stay on the same “page” they were on before the query was run. In other words, they’ll be sorted to the bottom of page one, not of the last page. Anyway… consider this most useful in cases where you’re setting posts_per_page to -1 or some arbitrarily large number (e.g. 999).

The function

This simple (and highly compact) function accepts a field name (and a boolean for whether or not it’s a custom field [meta data]), then takes the array of posts in the main query ($wp_query), splits them into two separate arrays — one with the non-empty values for your selected field, one with the empty values — and then merges those arrays back together, with all of the non-empty values first. (Other than shifting empties to the back, it retains the same post order from the original query.)

function sort_empty_last($field, $is_meta=false) {
  global $wp_query;
  if (!$wp_query->is_main_query()) { return; }
  $not_empty = $empty = array();
  foreach ((array)$wp_query->posts as $post) {
    $field_value = !empty($is_meta) ? get_post_meta($post->ID, $field) : $post->{$field};
    if (empty(implode((array)$field_value))) { $empty[] = $post; }
    else { $not_empty[] = $post; }
  }
  $wp_query->posts = array_merge($not_empty,$empty);
}

Calling the function

As I said, this function is designed to work directly on the main query. You just need to call the function right before if (have_posts()) in any archive template where you want it to apply. Because of the way it works — especially the posts_per_page consideration — I thought calling it directly in the template was the most clear-cut way to work with it. Here’s an example of the first few lines of a really basic archive template that uses it, looking for a custom field (meta data) called deadline:

<?php

get_header();

sort_empty_last('deadline', true);

if (have_posts()) {

Responsive horizontal scrolling tables for phones in CSS with no additional HTML

For some reason I always forget how to do this, and most tutorials out there suggest wrapping your table in a container <div> tag but as long as you’re using <tbody> (and, ideally, not using <thead> or <tfoot>) then it’s easy with a small bit of CSS and no HTML changes at all.

The trick is to display your <table> as a block, and your <tbody> as a table!

Here’s the code. Fit it into whichever breakpoint makes sense for your site. Mine shown here is the standard phone width breakpoint for WordPress (782 pixels):

@media screen and (max-width: 782px) {

  table {
    display: block;
    overflow-x: auto;
    width: 100%;
  }

  table tbody {
    display: table;
    width: auto;
  }

}

Of course, right after I posted this I found that I had already written about it two years ago with a slightly different solution. Never hurts to have a slightly different perspective.

Using The SEO Framework with Advanced Custom Fields

I’m going to go out on a limb and guess that I am not the only WordPress developer who in recent days (in the wake of their obnoxious Black Friday dashboard ad) has switched allegiance from Yoast to another SEO plugin, and that many of those who find themselves in a similar boat (to mix metaphors) have switched to The SEO Framework.

I’ve only been using it for a couple of days, but I already love it. It does all of the things I actually used Yoast for, without any of the other stuff I did not use it for. I mean honestly, maybe readability scores and “cornerstone content” do provide an SEO boost, but I barely understand how to use these tools, so good luck explaining them to my clients in a meaningful way. I suppose they’re more of a tool for full-time SEO consultants who need to pad out their billable hours. (Sorry not sorry. My opinion on the business of SEO hasn’t changed all that much since 2011.)

It wasn’t until the Black Friday ad that I really admitted to myself how much I don’t like Yoast. It does a lot of important things, and does them very well. But it’s obnoxious as hell about it. Pushing features you don’t really want or need into every page of the WordPress admin, and plastering its own over-designed admin screens with tons of garish ads promoting its “premium” features.

Yuck.

The SEO Framework encapsulates all of the key features I liked about Yoast into a single configuration screen, which kindly adheres to the standard WordPress admin UI design language instead of infusing its own brand style into every button and metabox border. It’s refreshingly boring to look at. And it just has the stuff I actually use, like title and description, OpenGraph tags, sitemap XML, the basic elements of SEO that unequivocally matter and can be a pain to build and maintain on your own.

But enough about all of its great features. There’s one key thing it lacks: support for Advanced Custom Fields. My standard “modular design” theme relies almost entirely on ACF’s Flexible Content fields to work its page layout magic, and with all of the page content stored in custom fields instead of post_content, there’s nothing for The SEO Framework to latch onto to auto-generate meta descriptions.

Fortunately, the developer has built in some hooks to allow you to customize the meta description output.

Here’s a barebones starting point:

function my_seo_framework_description($description, $args) {
  if (empty($description)) {
    $description = ''; // Add your own logic here!
  }
  return $description;
}
add_filter('the_seo_framework_custom_field_description', 'my_seo_framework_description', 10, 2);
add_filter('the_seo_framework_generated_description', 'my_seo_framework_description', 10, 2);
add_filter('the_seo_framework_fetched_description_excerpt', 'my_seo_framework_description', 10, 2);

As the developer notes, it’s very important for SEO not to just output the same static description text on every page. You need to have a function of your own that will read your ACF field content and generate something meaningful here.

Fortunately in my case, I had already done that, for generating custom excerpts from ACF content, so I was able to just stick a call to that function into the // Add your own logic here! line. You’ll need to customize your function to suit your specific content structure, but here’s the post that I used as a starting point for my function.

Have fun!

Why live shipping calculations for ecommerce are the bane of my existence

I’ll be frank: I don’t really like ecommerce. I’ve floated for most of my 23-year career in web development in the fantasy of a Platonic ideal of an online world: bytes streaming from point to point on a perfect network. No messy complexities of the physical world to concern me.

Well, obviously none of that is true. And I’ve been dealing with ecommerce in some form or another for at least 20 of those 23 years. But I still find the nitpicky details of sales tax and shipping logistics both mind-numbingly dull and infuriatingly (seemingly pointlessly) complex.

Which brings me to the crux of today’s rant. Bringing simplicity to all of this… or, at least, to the part I have to deal with.

Granted, I am not an accountant, and I don’t have to deal with balancing my clients’ numbers at the end of the quarter. But from my perspective, there are some simple solutions to the fact that calculating shipping sucks.

These days I deal mainly with WordPress, and therefore with WooCommerce. There are a lot of extensions to WooCommerce now for live shipping rate calculations. You can pay an annual fee for tools that connect you to shippers’ APIs to give you exact shipping rates. And all you need to do is:

a) determine exact weights and dimensions of all of your products,

b) figure out exactly what types of boxes/packages you’re going to use (and determine their exact weights and dimensions), and

c) plug all of those numbers into WooCommerce, and also

d) decide which of those types of packages are applicable to which products, create the appropriate Shipping Classes, and apply them to the corresponding products; then

e) set up Shipping Zones for each part of the world you’ll ship to,

f) add shipping methods in each of those zones for each of the shippers you’re working with,

g) go into each shipper, in each Shipping Zone, and select from the dozens of shipping services each shipper offers (although you might have to do this multiple times, if you want to restrict the methods you offer based on your Shipping Classes), then

h) test adding these products to your cart, enter an applicable shipping address, and see if you get the right shipping charges… or any shipping charges at all… because you very likely won’t, because for instance you may not realize that

i) WooCommerce Services no longer offers free access to USPS live shipping calculations for new sites created after WooCommerce 3.6, and you now need to buy a separate extension for $79/year to do that. But the free plugin won’t give you any indication that this is the case; it will still show you all of the shipping methods in the configuration pages, it just won’t give you back any rates on the checkout page, and you won’t have a clue as to why until you stumble upon a small note to this effect on the WooCommerce documentation website.

In case you didn’t guess, I just went through all of this today. And this was on a site that only has 7 SKUs!

Of course, there’s a different way to approach this. You could, instead:

a) offer flat-rate shipping.

Obviously it’s not quite that simple. But here’s the thing: there’s no law saying you can’t charge customers more for shipping than it costs you. That’s the whole idea behind “shipping and handling.” In fact, whether you’re doing the fulfillment yourself, or you’ve hired an outside fulfillment vendor, you’re paying more to ship those products than the shipper’s calculated rate anyway, so you should charge more.

Here’s what I recommend.

First, decide on a shipping method. Or maybe a few. Depending on what you sell, you might want to use USPS Media Mail (if it’s applicable). Otherwise, you’ll most likely want to offer two shipping methods: a) ground (cheap but slow) and b) 2-day air (more expensive but reasonably fast). You choose the vendor you like best: USPS, UPS, FedEx, some guy they call “Crusty Pete” with a rusty old pickup truck, etc.

Find out approximately what it will cost, on average, to ship each of your products, by each of these methods. (This is a good use for Shipping Classes.) Then tack on an arbitrary “handling” fee, say $1 per item for ground and $2 per item for 2-day air.

Now, decide: do you want customers to see a shipping charge, or do you want to offer “free” shipping? My logic for this is relatively easy. You just need to balance these two considerations:

People like to see free shipping.
People don’t like to see ridiculously high prices for individual items.

If you are going to offer “free” shipping, you’re really going to take that approximated shipping cost for each item and add it to the selling price of the item itself. Does that make the price of the item seem exorbitant? If not, go for it. People like free shipping.

As long as the shipping price itself isn’t so huge that it causes people to balk when they see it on the checkout page, you can keep the item prices low and present the shipping charge there. Or, a mix: offer “free” ground shipping, included in the price of the item, and then at checkout give the option of “free ground shipping” or paid 2-day air, with the 2-day air prices set at whatever you determined they should be, minus the amount you had already rolled into the item prices for free ground shipping.

Flat-rate shipping isn’t literally flat rate, as in one fixed price no matter what you’re buying. Generally it’s a formula… a base price, plus a multiplier based on the number of items.

For instance, your flat-rate formula might be $5 base plus $2 per item. The customer buys one item? $7 shipping. Two items? $9 shipping. You just plug the formula into the settings, and the rest gets calculated automatically.

The whole purpose of having shipping charges at all is to cover your expenses. Paying an annual license fee for a live shipping calculator extension, plus paying a developer or consultant for their time in helping you wade through all of these complicated configurations, and troubleshoot the problems that crop up, only to end up charging customers exactly what you’re paying for shipping and not a penny more just ends up costing you money. And the fact that these extensions include an option to add on an arbitrary amount above the calculated rate gives the game away… You don’t have to charge customers exactly what you’re paying for shipping. You just need to cover your expenses.

Your time and (my) mental wellbeing are expenses too. As long as your accountant doesn’t get too frustrated with the variable amount of ancillary income you’re getting from the modest overage you’ve calculated into shipping and handling charges, free or flat-rate shipping is a much simpler and more manageable alternative to live shipping calculations.

iCloud Drive: Don’t do what I did!

Being deeply immersed in the Apple ecosystem, a couple of years ago I made a decision:

I’ll move all of my work files onto iCloud Drive!

I work (as in, write code and edit image files) mainly on my Mac. But I was seduced by the possibility of accessing all of my work files in a pinch on my iPad (which I still had at the time) or even my iPhone. Plus, since my files would be in “The Cloud”, I could even access them from another computer (or from my Mac when booted into Windows) if I needed to, by logging into my iCloud account from a web browser.

It seemed… so obvious. So perfect.

Umm… maybe not.

For the past two years, I have been constantly fighting with iCloud Drive. One of its signature features is that it can manage disk use on your Mac automatically, so as your hard drive fills up, it deletes files you haven’t used in a while, keeping them safely in the cloud while freeing up disk space on your Mac. And with my MacBook Pro sporting a (meager?) 256 GB hard drive, with 40-odd GB allocated for a Windows partition, and over 60 GB occupied by Logic Pro X sound samples, my drive is filling up constantly.

While this is great in principle, it is completely unworkable in practice for three interrelated reasons:

If you have a large amount of data in play here (for me, it’s in the vicinity of 100 GB), iCloud Drive may get to a point where it is constantly transferring data. If you’re not on a gigabit fiber connection, this can both use up all of your Internet bandwidth and take ages.
Because #1 is taking place constantly, if you do find yourself needing to grab one of those files that has been deleted locally (as indicated in the Finder by a cloud icon with a down-pointing arrow), you may find yourself waiting several minutes for the file to become available, even if it’s small (as in, under 1 MB).
In an effort to make this all appear seamless to the user, the Finder represents cloud-only files as… regular files. But they’re actually just pointers with a hidden .icloud filename extension… as you’ll find if you ever try to perform Finder actions inside another program, such as syncing files to a web server using Transmit.

All of this might be tolerable if Apple gave you any control whatsoever over which files get deleted locally. But they don’t.

It gets worse.

What’s worse than getting stuck in this situation? Trying to get out of it. It’s like quicksand… the more you struggle against it, the faster and deeper you sink into it.

I made the decision earlier this week to extricate myself from the iCloud Drive nightmare, by buying a 250 GB SanDisk external SSD. First off, a little unpaid plug for this drive… it is awesome. It’s super light and small, seemingly at least as fast as the internal SSD in my MacBook Pro (in that I am able to transfer multiple gigabytes of data in seconds), and it even looks cool. I’m going to be using it all the time, so I’m actually considering putting adhesive Velcro on both it and the top of my MacBook Pro so I can keep it permanently attached. (Which says a lot about how much my regard for Apple has fallen lately — in the past I would never have sullied the exterior of an Apple laptop with something adhesive.)

So anyway, external SSD acquired, my goal was to start transferring my files from iCloud Drive over to the SSD.

Uh… good luck with that.

Because I’m now at a point where I have more than double the amount of data stored on iCloud Drive as I have available space locally, a majority of my files are now only in “The Cloud.” Ugh. Which means waiting for all of that data transfer stuff to happen. If only I could, somehow, bypass this broken process, I thought.

There has to be a way.

tl;dr Nope.

So, here’s the thing. I’ve been using iCloud Drive for the bulk of my cloud-based file storage, but I do use other services as well. I have Google Drive. I have Dropbox. I know how they work.

Specifically, I know that you can, y’know, like, select a folder and download the entire thing as a zip file.

So I thought to myself, I’ll just go to iCloud in a web browser and do that! Download the whole friggin’ thing as a big zip file, or maybe a few zip files, and be done with it.

Nope.

For whatever reason, iCloud doesn’t let you do that. Probably because of the whole seamless “It just works” Kool-Aid drinking song everyone in Apple land has been singing. (Myself included, mostly.)

You can only download individual files, not folders, from the iCloud web interface. It does let you select multiple files at once, but only within one folder.

Check out this delightful thread full of know-it-all asshats whose response to a legitimate question — why doesn’t Apple allow this thing that every competing service does? — is to challenge the validity of the question and the intelligence of the questioner. (That thread is now closed so I’ve just opened my own new thread on the topic. Watch this space for trolls!)

There. Are. Plenty. Of. Reasons. A. Person. Might. Have. For. Needing. To. Do. Something. That. You. Have. Not. Previously. Considered. Stop challenging their premises and answer their question, or shut the hell up.

Whew.

I ended up “solving” the problem by resigning myself to the fact that it wouldn’t be completely solved. So instead I took the drastic approach of temporarily logging out of iCloud completely, just so I could strand the files I did have saved locally, and copied them to the SSD.

Then I logged back into iCloud Drive and tried to get it to stop syncing my files by unchecking the Desktop and Documents Folders option.

The only problem is, I didn’t have my work files in those folders. I had them in a separate top-level folder in the iCloud Drive that I created myself. Because, you know, you can do that and it didn’t seem like a crazy idea or anything.

It was.

I discovered this morning that even though I had done all of this and tried to purge the nightmare of constant iCloud Drive syncing from my Mac life, once I had logged back into iCloud, the Mac went right back to quietly, constantly, syncing that iCloud Drive data on my Mac. As I type this, I have a Finder window open to my iCloud Drive and in the status bar it says “downloading 120,079 items (36.14 GB of 48.66 GB)”. Fun!

So, my new plan for today is to watch that window, and as the little cloud icons next to individual folders goes away, I’m copying those folders to my SSD and then deleting them from iCloud. My assumption is that as I do this, I am freeing up more local space and iCloud will continue to download the remaining items, and eventually I’ll have everything transferred over.

But please… do yourself a favor and don’t do what I did. iCloud Drive is not suitable for professional use.