Speaker Interview: Louise Grandjonc

Cleaning out crocodiles teeth with PostgreSQL indexes - a story on all the index types in PG   Thursday 11:50   Casablanca

Twitter: @louisemeta Blog: louisemeta.com LinkedIn: louise-grandjonc Company website: www.citusdata.com

Could you briefly introduce yourself?

I’m Louise, I work at citus data in the solutions engineering team. Previously I was a lead developer in a crowdfunding company. I really like writing python and SQL.

How do you engage with the PostgreSQL Community?

I try to attend as many events as possible, I talk at some, and the rest of the time I listen.

I’m also in the PostgresWomen group, our goal being to make women more visible in the community and hopefully inspire others to work with databases, and feel welcomed. It’s part of the reason I am a speaker, speaking in public is terrifying, but it’s important to show that we exist (unlike unicorns, unicorns don’t exist, it’s a lie).

Have you enjoyed previous pgconf.eu or FOSDEM conferences, either as attendee or as speaker?

I have been to two pgconf.eu and attended FOSDEM three times, every time as attendee.

What will your talk be about, exactly? Why this topic?

My talk is about the internal data structure of indexes.

It all started when I was working with fulltext search and postgres. I was wondering what that GIN thing had different from a BTree. So I decided to submit a talk and figure it out later.

What I really wanted was to go quite deep into the algorithms used to search, insert, delete from an index. Which is why I read the source code and part of my talk is going into the implementation of indexes in postgres.

What is the audience for your talk?

The audience I’m trying to reach is beginner to intermediate.

What existing knowledge should the attendee have?

Well, you don’t even have to know what an index is, as I will explain. However, it will definitely help to have some experience with dealing with indexes.

What is the one feature in PostgreSQL 11 which you like most?

I really like the possibility to use INCLUDE in unique indexes to have columns that are not part of the unique constraint but can be used for index scans. This is fun.

And also, I’ve been waiting for years for the compile-time option ALLOW_DANGEROUS_LO_FUNCTIONS to be removed.

Which other talk at this year’s conference would you like to see?

Also, I apparently have to attend my own talk.