RStudio Data Viewer - Very slow vs Stata and SAS

Hi all,

I have been using R, SAS and Stata since 2008. I'm a big fan of RStudio and R in particular except for the data viewer in Rstudio. It's 15+ times slower in scrolling horizontally or vertically as compared to SAS and Stata with the same data file making the concept of the data viewer in RStudio unusable. It's with all datasets. Right now I'm working on a small dataset of 2000 variables and about 140k obs and RStudio freezes for 10 seconds when i open the viewer by clicking on the dataset icon in the Environment. The viewer reacts to the horizontal scrolling bar clicks with about 2-3 seconds delay and it's impossible to drag the slider horizontally at all. I can post a recording of a comparison between the 3 using some public data.

It's been the case for all RStudio versions for all R versions for the last 6-7 years. Viewing and scrolling the same data in SAS with its "data on disk" and STATA with its "data in ram" is instantaneous. My system is AMD TR 1900x at 4ghz with 128gb of DDR4 2733 cas 14 ram.

Is there some setting I'm missing or is there a different package for viewing data in RStudio? Right now I have to save the data in either dta or sas7bdat for viewing.

Thank you,

E

Hi,

It's interesting that there are no replies to this. Is this too common or too uncommon? Are there any ways to optimize the data viewer and make it workable? I will try create a side by side comparison video next week.

Thank you,

E

I use R every day I'm at work, and somedays when I'm not as a nerdish hobby, and the times I use the data viewer are ... never.
if I want to understand the content of a data.frame , I will use code in the console to get me the level of description I require. I don't know how common or uncommon I am, but if there are others like me, this might explain the response (or lack) somewhat.

It's the polar opposite in my department, data viewers are used all the time along with all the description commands/code for the summaries and tables. It's a tough sell to introduce RStudio to my colleagues when they can't even view the data efficiently. So nobody from R users uses the RStudio data viewer? I assume there is no roadmap to improve the viewer at least to the level of Stata?

Btw, thank you for replying to the post! Any type of feedback/info is appreciated.

It is also my experience that the Viewer gets slow scrolling "big" tables. Similarly, I tend to explore my data first in the command line (displaying it with as_tibble() or head()), or, when I'm interested in a subset of records, to filter it beforehand:

df |>
  filter(some_id == "some_value") |>
  View()

The search box in the Viewer is also pretty fast.

If scrolling speed is really a concern, you can try this:

modified_df <- edit(df)

The result will depend on the system. On Windows, it opens a basic, slightly ugly but really fast editor. It's the same viewer that appears if you use View() without RStudio. On other systems, it might open an installed text editor, which might be less practical.

Note that this command is meant to edit the data, so after running it returns the modified data.frame. You probably want to assign it into a variable, otherwise it all gets printed in the console.

1 Like

This is great feedback, thank you. We're aware of a number of issues in the Data Viewer, and we will be undertaking a complete refactoring of it in the near term (currently scheduled for a fall release). Known data viewer issues: Issues · rstudio/rstudio · GitHub

2 Likes

Thank you, @MikeBess! This is encouraging!

This topic was automatically closed 21 days after the last reply. New replies are no longer allowed.

If you have a query related to it or one of the replies, start a new topic and refer back with a link.