Recent

Author Topic: StackOverflow 2022 Developer Survey  (Read 798 times)

avra

  • Hero Member
  • *****
  • Posts: 2368
    • Additional info
StackOverflow 2022 Developer Survey
« on: July 11, 2022, 11:11:51 am »
https://survey.stackoverflow.co/2022/#technology-most-popular-technologies

It was commented on Marco Cantu's blog:
https://blog.marcocantu.com/blog/2022-june-delphi-stackoverflow-survey.html

Please look at it just as an interesting to read survey stats, and not like a reason for another passionate anti TIOBE like flame war.

For me, regarding databases the most interesting part was that PostgreSQL was number 1 choice for professional developers (pleasant surprise), and that InfluxDB was not on the list at all (unpleasant surprise).
ct2laz - Conversion between Lazarus and CodeTyphon
bithelpers - Bit manipulation for standard types
pasettimino - Siemens S7 PLC lib

PierceNg

  • Full Member
  • ***
  • Posts: 199
Re: StackOverflow 2022 Developer Survey
« Reply #1 on: July 11, 2022, 01:42:25 pm »
For me, regarding databases the most interesting part was that PostgreSQL was number 1 choice for professional developers (pleasant surprise), and that InfluxDB was not on the list at all (unpleasant surprise).

For monitoring I run the TIG stack: Telegraf, InfluxDB and Grafana. When the time comes to reinstall, I'll most likely migrate from InfluxDB to some time series DB extension for PostgreSQL. The advantages are familiarity and size of the PostgreSQL ecosystem.

Marc

  • Administrator
  • Hero Member
  • *
  • Posts: 2556
Re: StackOverflow 2022 Developer Survey
« Reply #2 on: July 12, 2022, 10:28:40 am »
I'll most likely migrate from InfluxDB to some time series DB extension for PostgreSQL.
My use case benefits from being able to set InfluxDB data to automatically expire after some time period (like 1 year). PostgreSQL can not do that and I would have to schedule some long SQL query and use CPU intensively while executing that query. InfluxDB simply deletes the file with old data and continues without interruption and extensive CPU load.

In PostgreSQL, if you partition your table on a time range you can drop a partition instantaneously. If you set the range for instance on a day, you can drop your data for given day without heavy queries, with little to no overhead
//--
{$I stdsig.inc}
//-I still can't read someones mind
//-Bugs reported here will be forgotten. Use the bug tracker

avra

  • Hero Member
  • *****
  • Posts: 2368
    • Additional info
Re: StackOverflow 2022 Developer Survey
« Reply #3 on: July 12, 2022, 01:32:16 pm »
In PostgreSQL, if you partition your table on a time range you can drop a partition instantaneously. If you set the range for instance on a day, you can drop your data for given day without heavy queries, with little to no overhead
Thanks Marc for this correction. I was not aware of that. According to documentation, you would steel need to manually drop or add date range partitions with PostgreSQL. However it seams to be much faster then bulk delete so it could be useful in some use cases.
« Last Edit: July 12, 2022, 01:35:24 pm by avra »
ct2laz - Conversion between Lazarus and CodeTyphon
bithelpers - Bit manipulation for standard types
pasettimino - Siemens S7 PLC lib

Marc

  • Administrator
  • Hero Member
  • *
  • Posts: 2556
Re: StackOverflow 2022 Developer Survey
« Reply #4 on: July 13, 2022, 01:50:08 pm »
Thanks Marc for this correction. I was not aware of that.
I can imagine. At the time this got introduced, we just had finished our own solution with daily tables  :(
Quote
According to documentation, you would steel need to manually drop or add date range partitions with PostgreSQL.
There exist a package which can handle the auto add and drop. OTOH, if you have some job running which drops a partition every month, the same job can add a new month
//--
{$I stdsig.inc}
//-I still can't read someones mind
//-Bugs reported here will be forgotten. Use the bug tracker

avra

  • Hero Member
  • *****
  • Posts: 2368
    • Additional info
Re: StackOverflow 2022 Developer Survey
« Reply #5 on: August 19, 2022, 03:57:37 pm »
I really do like PostgreSQL as a general SQL database, but for time series data some specialized database like InfluxDB hits more checkmarks in my use case, and it even has SQL like syntax - so it is easy to remember.
I have just bumped into CrateDB, which claims to be time series database compatible to PostgreSQL wire protocol, has standard SQL, and in some use cases much faster then PostgreSQL.

Sounds too good to be true...  ;)

https://crate.io/products/cratedb
https://crate.io/use-cases/time-series
https://crate.io/docs/crate/reference/en/5.0/interfaces/postgres.html
ct2laz - Conversion between Lazarus and CodeTyphon
bithelpers - Bit manipulation for standard types
pasettimino - Siemens S7 PLC lib

 

TinyPortal © 2005-2018