F2023 #08 - B+Tree Indexes (CMU Intro to Database Systems)

[@potapow8684]

Timestamps 20:27 - B-Tree vs B+Tree 23:03 - B+Tree Insert 26:23 - B+Tree Delete 33:37 - Select conditions 40:50 - Duplicate keys 46:05 - Clustered indexes 50:37 - Node size 52:38 - Variable length keys 55:29 - Intra-node search 59:27 - Optimizations: Prefix compression 1:00:21 - Optimizations: Deduplication 1:01:47 - Optimizations: Suffix truncation 1:02:48 - Optimizations: Pointer swizzling 1:06:07 - Optimizations: Bulk insert 1:08:00 - Write-optimized B+Tree 1:12:43 - Demo (BTree vs Hash index)

[@kartikmahajan4405]

hey thanks for these timestamps.

[@llight1635]

great course

[@changliu9090]

How should I delete the leaf node if the key in leaf node is also in some internal nodes?

[@nanunsaram]

Darn great!

[@blockpaper]

Didn't understand problem with duplicates in tree and how appending record_id solves it. Moreover, when I want to delete on 6, how I (DB systems) know what is record_id? I thought point of index is giving you record_id as a value so DB system can fetch tuple using record_id. When we use record_id as a part of a key what is the point of index then?

[@Jael860]

When you are doing a deletion, what you are actually doing is a deletion of a record from a table. Then in order to keep the index in sync with the table, you have to do a deletion on the index (the b+tree). Since you started by doing a deletion of a record, you know the record id, therefore you can compute the key needed to do the deletion in the B+tree.

[@NamYounghwan-n1e]

thanks a lot!

[@yashthakkar4499]

i am going to go with bushy tree lol.

[@WangMiracle]

👏

[@tsunningwah3471]

i wanna diwe

[@tsunningwah3471]

fi

[@tsunningwah3471]

zha