Create Your Own Database from the Ground Up - Part 1 - Balanced Binary Trees

[@HanCurunyr]

As a DBA with 10 years of experience, that was the most friendly explanation of inner workings of a non clustered index that I ever saw, great work Gary, I'll link this video to every newbie that needs to understand indexes

[@endoflevelboss]

As a DBA with 10 years experience? So quite a noob really on the professional ladder...

[@HanCurunyr]

@@endoflevelboss boohoo

[@endoflevelboss]

​@@HanCurunyrmakes zero sense. That will be the 10 years inexperience talking...

[@rodrodrigues1729]

@@endoflevelboss true... I have 20 years in the field and I'm pretty sure I know nothing.

[@thepresi2]

what a serendipitous post! I'm working on a database project written in C and I was wondering what an ideal course of action would be. Thanks Gary!

[@lavisharma3210]

Thank you so much Gary, this is exactly what I was searching for, I am planning to work on a side project for myself, creating a database in golang and have been researching about databases and their internal workings and parts (mysql to be specific). This series came at the right time for me. Thanks again !

[@GaryExplains]

Glad it was helpful!

[@johng7rwf419]

Thanks Gary, very interesting. Takes me back to the 80's where I was writing databases and optimization routines for surface mount placement machines. J

[@danielgospodinow]

Great material! Thanks a lot, Gary!

[@willthecat3861]

Interested in this for cataloging all equipment in different buildings, and locations... for finding it when they need it.

[@noodlish]

Really cool. Did you consider using a skip list instead? I find those fascinating due to their random nature.

[@GaryExplains]

I didn't consider that specifically, however I agree that there are lots of ways to do this.

[@m8hackr60]

Hi Gary, I find this very cool and I might have to run some performance test on this one day. :) But really fantastic work. This level of proficiency typically takes a well organized and dedicated team. This coming from a DB Engineer of 18 years (in IT for 25) for a SaaS company working with Redis, Cassandra, PostgreSQL, MSSQL, and Oracle... among the top.

[@GaryExplains]

Very kind of you, thanks.

[@jaydeep-p]

I love the video, while watching it I was thinking where a trie would fit into this. It would work just as well for search and insert operations right?

[@GaryExplains]

Interesting idea. I have never implemented a trie, but according to my quick bit of research the search time for a trie is O(n) compared to O(log n) for an AVL tree. But I could be wrong!

[@jaydeep-p]

@@GaryExplains It would be O(m) where m is the length of the longest key.

[@akiratoriyama1320]

Super interesting!!!!

[@joeedh]

I've always wondered why people dont use btrees (block trees) for in memory stores. DRAM is still pretty slow, dunno why everyone assumes the optimal child number is 2.

[@MichaelSkinner-e9j]

I need to get cracking on Nvidia‘s courses (Thankfully they come included on the Tegra boards) but I keep wondering how much of a database (SD Card size) would be needed on flight pathing and controls, While also allowing a user interface that’s seamless and easy

[@muddyexport5639]

Good job & thanks!

[@GaryExplains]

Thank you too!

[@tonysheerness2427]

I understand the balance part but it puts it out of Alphabetical order, how does it manage it?

[@HanCurunyr]

This is how a non clustered index works in SQL Server, it sorts the data, in case of strings, by alphabetical orders, checks for the middle element, sets it as a root and starts balancing the tree from there, as time passes, new data is inserted, deleted or updated, the tree becomes unbalanced, or fragmented as the SQL Server calls is, if the unbalance factor is less tan 30%, its possible to just reorganize the leafs, but if its more that 30%, we usually rebuild the index, as we delete the entire tree and start from scrath again

[@QuikRay]

You create data bases with windows Access or Libre Office

[@Maruman_man]

This is at the machine level. You are describing application/end user apps

[@MaximoJoshua]

That JSON in the beginning is going to be invalid because the last quote is a "curly" quote, and doesn't match the opening one which is a straight quote...the word price is quoted like this "price“. Really distracting was all lol.

[@GaryExplains]

🤦‍♂️

[@MaximoJoshua]

@@GaryExplains Great video though, enjoyed it a lot.

[@aizensosuke4660]

Number 1😂😂😂

[@igaviotis]

Sorry Gary, but what you describe is NOT a database, it is the implementation of the abstract data type called dictionary. A database cannot assume that all data resides in memory (that's a data structure like the balanced tree you implemented). Just saving and loading data in storage is barely sufficient for a db. I would argue that a dbms not offering schema creation is hardly a db anyways..

[@GaryExplains]

Fair enough, your criteria are different to mine. I respect that. I guess you also don't like any of the nosql type databases or key-value stores.

[@mintoo2cool]

why is everyone all of a sudden trying to create databases ‘from ground up’ ?

[@GaryExplains]

All of a sudden? Has there been a mad rush to create databases? I must have missed that 😜 These videos (in fact most of my videos) are about teaching things, explaining things, it is even in the name of the channel. Discovering how to create a database from scratch is called learning.

[@mintoo2cool]

@@GaryExplains Don't get me wrong. I appreciate your videos .. I blame the youtube algorithm more than you for inundating my feed with a bunch of these ... I guess I should have shut up and clicked off the video instead of getting short in the comment section ... because it's not something I wanted to watch ... feel free to remove this comment thread if it takes away from the video. Nevertheless, I appreciate your videos in general. Thanks for all your hardwork in educating us folks.

[@wandiletembe]

👨🏿‍💻🍿🇿🇦