Vscode-jupyter: Could each cell have only one corresponded output tab on the Python Interactive Window?

[AeneasZhu]

Each time I run a cell on the left side, the output on the right side (python interactive window) will create a new tab each time. But when I run the same cell, unlike in jupyter notebook , the output window will also create a new output tab too. In jupyter Notebook, each cell has only one single corresponded output tab. But in VSCode, each cell could have many output tab on the right side, if I run the same cell many times.

As shown by the image, one cell(marked by the red square) has more than one output tab on the right window. This is very unconvenient when you have to check the output above again

It is very unconvenient sometimes, when you have to check the output shown above by using mouse (I use Vim, and I hate using mouse when coding).

At last, I also recommed the right side (python interactive window) will scroll following the left side's scrolling by mouse or vim, just like the markdown preview will scroll when scroll on the editor side.

[StevenLi-DS]

@AeneasZhu which theme are you using buddy?

[AeneasZhu]

@AeneasZhu which theme are you using buddy?

Quiet Light. But I have installed 'Winter is coming'. I don't know whether you can see that without 'Winter is coming'

[rchiodo]

This is actually by design at the moment. The interactive window is not the same as a jupyter notebook. It's more like a console or like spyder where it keeps a 'history' of everything that ran.

There are other ideas we've been talking about for editing in more of a jupyter style though:

1. Keep the interactive window in sync with your file (like you suggested above)
2. Have an 'inline' mode where the output for your cells ends up inside of the editor.

We'd love to hear if you have any other ideas.

[gramster]

1. Like 1, but remove the old output (wherever it is) and add a new output on end of history - so history is incomplete but has outputs ordered by execution recency. This is a hybrid of the current approach and the Jupyter approach.
2. Like 3, but using dependency analysis to see if removing the old output would be removing an important intermediate computation, and in that case retaining the old output too. That means that the output should still be a reproducible execution sequence; we only prune old outputs if they are redundant.

[StevenLi-DS]

@AeneasZhu I think I meant to ask about the font family.

[AeneasZhu]

@AeneasZhu I think I meant to ask about the font family.

Comic Sans MS

[greazer]

This is further info for dealing with microsoft/vscode-python#4575.