The reviews are in for our “Modern Search” series, covering everything you need to know to be effective with full-text search. And we have one big episode left in the series. 😯 https://t.co/jl8DCVCEIw

Point-Free's (@pointfreeco) episode on SQLite full-text search is 🔥. Highlighting matches and generating snippets might seem basic, but getting it right is what separates good apps from great ones.

SQLite's full-text search module provides 'snippet' and 'highlight' functions for bolding the search matches and focusing in on just the matches of a large document. In the screenshot below, the leading characters of the notes are trimmed because the search terms match at the

Apple’s not the only one with a big upcoming release!

SQLite’s full-text search tools come with many bells and whistles, including support for highlighting search terms in your UI, as well as generating snippets for where matches appear in a larger corpus. Let's explore each of these tools in-depth. 👉 https://t.co/8yITibv6ca

Let’s take advantage of the superpowers of full-text search, starting with relevancy. We will explore the ranking algorithm of SQLite’s FTS5 module, see how documents are scored, and how we can tweak the scores based on the column containing a match. https://t.co/gd7Uipsbvv

We can customize how full-text search weighs your data to fine-tune searches. For example, we can make matches in the title of a reminder count for more than matches in the notes, and make notes count for more than matches in the tags. Learn how: https://t.co/gd7UipsJl3

This week is where full-text search really shines. It sorts matches by relevancy based on many factors, such as frequency and position of matches, document length, and more. We get access to all of that with just one simple SQL clause: ORDER BY rank 🤯

Calling Swift functions from SQLite is incredibly powerful, and we've just made it easier than ever with a brand new @‌DatabaseFunction macro. Get all the details here: https://t.co/WKJ5o5WCAv

Next week we explore how to fine-tune some of the more advanced aspects of full-text searching, such as results ranking. This will allow us to scores matches in a reminder's title a little higher than in its notes or tags. Until then, catch up here: https://t.co/axpzsO12T4

A common refrain we hear in the community is “you shouldn’t depend on a library for your architecture”. We have no interest in forcing anyone to use our libraries, but we do feel there is nuance worth considering. https://t.co/PR1ym3Qtza

This is a great tutorial for getting started with SharingGRDB! Here are a few tips we noticed from going through the video in case folks are interested in a few features that could take their SharingGRDB to the next level (1/5)

Full-text search is a super power of SQLite, but is yet another thing hidden from you in SwiftData, even though under the hood it uses SQLite. This week we create virtual tables necessary to compute an FTS index, and then effortlessly search our tables. https://t.co/axpzsO12T4

This week we add some bells and whistles to the search feature of our reminders app, including debouncing database queries, adding a count query to display the number of completed reminders, searching tags, and improving database performance. https://t.co/y6cJiZRi5W

@MikelCobian @CarlosBBuild Too bad knowing that SwiftData uses SQLite under the hood doesn't actually give you access to SQLite's wonderful tools. If anything it's only sad to know this fact because SwiftData hides everything great about SQL from you ☹️

Our SQLite+CloudKit beta has been running for a few weeks, and we are hard at work putting on the finishing touches. Do you have any questions? Ask us! Either here or in the repo discussions: https://t.co/lk4pdIbbik We would love to get feedback during this period.

We recommend writing migrations as SQL strings. This gives you the full power of SQL, the most flexibility, and none of the drawbacks of magical, “lightweight” migrations of SwiftData: More details here: https://t.co/B7dJcPoQUC

Here is another example. Something as simple as adding a unique constraint to a column is just a few lines of SQL, yet dozens of lines of obtuse steps in SwiftData. Compare the SQL code on the left to the SwiftData code on the right: https://t.co/fMpfTFO37U

• Seamlessly search reminders and their tags with type-safe and schema-safe SQL • Aggregate counts computed right alongside search results • Automatic animations when search results change Are you dreaming? Nope. This is easily attainable using our SQLite tools 😊

Isn't this beautiful? A straightforward SQL query allows us to search reminders by their title, notes *and* tags via a many-to-many association. This is what SQL excels out, and why one should never put abstractions in place that hide it from you. Learn more in 🧵