Month: December 2018

“Schools are places where you send your children to be brought up by other children.”

– Alan Watts

Almost every answer on Stack Overflow for this is terrible on moderately complex or large real data – https://stackoverflow.com/questions/4329396/mysql-select-10-random-rows-from-600k-rows-fast.

Bad: ORDER BY RAND() LIMIT N

This means executing your query for the entire resultset and then ordering it and then chopping off the number you need. This can have truly dire performance.

Bad: SELECT MAX(id) FROM table
Don’t pick random numbers from 1 to MAX(id) – deleted rows will be null or result in you getting less rows than you want. Who says your id is even sequential/numeric?

OK: SELECT id FROM table

Then pick N at random (removing those already chosen if you don’t want duplicates) from the resultset in your chosen programming language.

If you just want one row from the db you can do this in SQL only as shown here: https://stackoverflow.com/a/31066058/375262. Doing N unique rows this way is left as an exercise for the reader.

If you have a gargantuan resultset even SELECT COUNT(*) might be slow. What would you do then?

Backticks around the reserved word.

I found this working with JOOQ’s implementation of the SQL IN predicate. You can’t write myJavaLibObject.in(…) in Kotlin. You’ll get “Element expected” because in is a reserved word in Kotlin. Instead you need backticks: myJavaLibObject.`in`(…) will do the trick.

Hard to believe but I think this is the simplest you can manage:

long millisSinceEpoch = timestamp.getTime();
Instant instant = Instant.ofEpochMilli(millisSinceEpoch)
OffsetDateTime dt = OffsetDateTime.ofInstant(instant, ZoneId.of("UTC"))

Yikes.

(Based on Stack Overflow tag count and upward trend as proportion of Stack Overflow questions. So it might just be the 50 most difficult-to-learn rising programming technologies!)

The List

1. Python

   Interpreted high-level programming language for general-purpose programming

2. React

   JavaScript library for building user interfaces

3. Laravel

   PHP web framework (“for web artisans”)

4. Pandas

   Data structures and data analysis tools for Python

5. TypeScript

   Superset of JavaScript that adds static typing (“JavaScript that scales”)

6. Amazon Web Services

   On-demand cloud computing platform

7. API

   The interface between two programs.

8. Azure

   Cloud computing service

9. Powershell

   Commandline shell and associated scripting language

10. Firebase

    Mobile and web application development platform

11. Selenium
12. Spring Boot
13. Docker
14. React Native
15. DataFrame
16. Unity 3D
17. Elasticsearch
18. Matplotlib
19. Go
20. Jenkins
21. Selenium Web Driver
22. Gradle
23. Machine Learning
24. Amazon S3
25. vue.js
26. ggplot2
27. Flask
28. ASP.NET Core
29. npm
30. Webpack
31. Mongoose
32. tkinter
33. Google Apps Script
34. Web Scraping
35. Spring Security
36. filter
37. https
38. Woo Commerce
39. Xamarin Forms
40. Web Socket
41. Android Recycler View
42. Kotlin
43. Redux
44. Google Sheets
45. Excel Formula
46. SASS
47. Hive
48. Java 8
49. Redis
50. CMake

The Top Ten

Chart of the top 10:

The Top Ten Without Python

Python dwarfs everything else so here’s a look without Python:

The Nearly Men

These tags were eliminated from the list solely on the basis of a 2018-only downward trend: R (would have been 2nd), Node.js (2nd), PostgreSQL (4th), numpy (12th), Express (14th), Apache Spark (14th), Tensorflow (18th), nginx (20th), Github (21st), Amazon EC2 (31st), ECMAScript 6 (39th), ffmpeg (46th)

Programming Languages That Make the List

1. Python
2. TypeScript
3. Go
4. Kotlin

A Rising Python Lifts All Python Libraries

python-3.x actually makes second place on the list but I rolled it into Python rather than make a redundant entry.

Django and Django Models were eliminated from the list despite being on an upward trend because they have not yet exceeded their previous peak in 2010. This resurgence, Pandas in fourth place and the presence of tkinter on the list speaks to the general rising of Python.

Methodology

Load all tags on StackOverflow by count descending.

Put each of them into StackOverflow Trends and judge by eye if they are currently trending up.

Where Do These Technologies Sit in the Overall List?

In the whole list of tags by count Python is sixth overall, Firebase (10th place in this list) is 90th, and CMake (50th) is 442nd.