Hacker News Reader

Comment by hyperbovine | Hacker News Reader

R is a but of a horror show under the hood, I agree, but if you're just an end user doing data analysis, consider:

    flights.iloc[0:10, flights.columns.get_indexer(['year', 'month', 'day'])])

versus

    flights %>% select("year", "month", "day") %>% head(10)

I could go on...

View on HN

You could have done

    flights[['year', 'month', 'day']].head(10)

which is not so different from standard R

    head(flights[c("year", "month", "day")], 10)

but it's true that the following may be nicer

    flights[1:10, c("year", "month", "day")]

(by the way using head(10) is not the same as indexing 1:10 if there are less than 10 rows)

I totally agree that it is a very efficient and powerful tool for ad-hoc data analysis. It's just not what I would view as a responsible choice for production / publication code.

    flights[["year", "month", "day"]].head(10)

It’s pretty sad that both are worse than

     SELECT year, month, day
       FROM flights
      LIMIT 10

SELECT before FROM isn't really a good thing.

That's a very dishonest example. Yes, terribly written pandas code looks terrible.

flights[1:10, .(year, month, day)]

for the data.table fans

Which is arguably the superior way to handle tabular data in 2020.

Reply on news.ycombinator.com