Tonight we write our first AQL graph query to display file names of ZIP attachments in the email view! #php #symfony #arangodb join us at https://bonn.live (or https://twitch.tv/larsborn if that's more your jam)!

AQL stands for Arango Query Language and is the "SQL equivalent" for the ArangoDb multi-model database software. At the end of last's stream, ChatGPT (3.5?) just spat out a working draft and today we try to reach parity in understanding AQL.

See you on the tubes!

Devrel fail of the day.

#mongodb disables comments on an otherwise, mostly-ok video on design patterns for document stores.

Community feedback would have been great here. I know most of the information applies to #Arangodb which I use.

https://youtu.be/bxw1AkH2aM4

#devrel #mongo

Well, I certainly am glad I didn't trust #ArangoDB to store four years worth of cached data from Twitter using their formula for #Docker ...

Oh, wait. I did do that.

And a restart of Docker just ate it ...

Got arangoimport loading 1M records from a flat EDI text file (CSV with 40 columns basically), and it finished in 8 seconds. 125k/s is very respectable. #arangodb

Still torturing a local deployment of #Arangodb as part of tech evaluation.

Yesterday got a graph building query syntax wrong, wanted to create just 1:1 edges between to document sets, instead it created N*M edges! Well, the machine ran out of disk and the underlying RocksDB halted. I killed the query and it left behind a graph with 2M edges. Deleting edges freed up all disk space.

Today been playing with full text search engine; you have to create views AOT, but it has all the knobs to tweak.

"We developed a #Lisp-like intermediate representation to be able to transport programs into the existing Pregel implementation in #ArangoDB. These programs are executed using the interpreter inside the AIR Pregel algorithm."

https://www.arangodb.com/docs/stable/graph-analytics-custom-pregel.html

This whole rant was brought to you by two recent experiences:

1) I am currently evaluating #Arangodb and find it delightful so far, but there's a huge discrepancy between it's raw performance when interacting with it directly, vs what I get out of the Java and Golang drivers.

2) I wrote a Spring Data Repository abstraction that is tenant-aware, every call of the Crud and JPA repositories take an extra tenant ID, because it's used in a #SaaS app. But I can't be trust it anymore. Too much magic!

Seeded Arangodb with 1B documents, running on my aging MacBook, and it was snappy!

Each "document" has an enforced JSON schema of (id: String, name: String, price: Double), plus a few random fields for each document.

Built-in UI suffices for learning & local admin, tons of metrics and timing hints. Good UX with syntax highlighting, sample queries.

Document lookup, graph relationship queries, & full text all work great.

Skip the Spring-Data driver.

Next we test operational stuff.

#arangodb

You might not need to code much for domain modeling. If you want to write a few entity descriptions using a pseduo-SQL, sketch executable functions in a pseduo-JavaScript, and enjoy free rest endpoints, repl, object graph, document store, full text search and CRUD operations: just stock #Arangodb console will do.

You will be interacting with a web UI, backed by a truly multi-model database that can persist data.

Don't trust it for prod, yet, but it's a perfect mind tool. #softwarearchitecture

Just some additional #introduction. My preferred environment at the moment is something like
- #rockylinux
- #flask
- #arangoDB (I like arango-orm for modelling)
- #htmx (extra exclamation marks 😉) to reduce or substitute js/js-frameworks.

Nevertheless I still do a lot of classic #Zope and #mongoDB stuff.

@skobkin @cauf@mstdn.social устали от капризности elastic и постоянных граблей. Используем для анализа большого объёма логов и threat hunting.
Хотели schemaless - просто json складывать, внятный язык запросов, не java и не потерять в скорости.
Выбрали https://www.arangodb.com
C++, есть графы, удобная админка, очень быстрая, если про индексы не забывать.
Очень довольны, но приколы тоже есть.
#bigdata #arangodb

RT @arangodb@twitter.com

Woop, woop! ArangoDB 3.5 is GA – Distributed JOINs, Streaming Transactions, extended #SearchEngine & #GraphDatabase capabilities.

Check out all shiny new features on our blog:

http://bit.ly/2KVtYCO

#NoSQL #MultiModel #ArangoDB #shinynewthings

🐦🔗: https://twitter.com/arangodb/status/1164158624253579265

RT @arangodb@twitter.com

New Case Study!
Meet @kode_solutions@twitter.com: Semantic Tweet #Analytics with #ArangoDB & #Rlang

Big thanks to @kode_solutions@twitter.com for developing the driver & sharing the case study! 🥑

👉http://bit.ly/2JNVjWZ
#NoSQL #Database

🐦🔗: https://twitter.com/arangodb/status/1151945546229309440

The whole open (and free) source infrastructure around Microsofts development-tools -- the, for me new, language-server microcosmos -- entering my #emacs workflow is thrilling. I have lingering doubts, remnants of the FUD days of yonder, but cannot see how MS could take the results of the adoption away from the FOSS community.

Check out lsp-mode, dap-mode and the servers available -- I use it for serverside js (#arangodb #foxx apps ) and PHP development