Spyder: View command to open DataFrame in Variable explorer

[Diego-MX]

Description

I come from R with RStudio and I'd like to integrate more Python in my workflow.
With Spyder, I have been able to make this easy. What I've found in Spyder + pandas is really cool and has helped me a bunch.
One feature that I think is comfortable from my previous workbench is the View command, perhaps explore makes more sense in Spyder environment... because of the Variable Explorer.
Say you have a table called my_df, execute
In [7]: explore(my_df)
in the command line, instead of clicking on it in the Variable Explorer. Thus allowing your hands on the keyboard =)

I didn't find a similar command here, so I'm taking the liberty to suggest it.
Cheers.

Versions and main components

• Spyder Version: 3.2.1
• Python Version: 3.6.1
• Qt Version: 5.6.2
• PyQt Version: 5.6
• Operating system: Linux-Ubuntu16.04

Dependencies

IPython >=4.0 : 5.3.0 (OK)
cython >=0.21 : 0.25.2 (OK)
jedi >=0.9.0 : 0.10.2 (OK)
nbconvert >=4.0 : 5.1.1 (OK)
numpy >=1.7 : 1.12.1 (OK)
pandas >=0.13.1 : 0.20.1 (OK)
psutil >=0.3 : 5.2.2 (OK)
pycodestyle >=2.3: 2.3.1 (OK)
pyflakes >=0.6.0 : 1.5.0 (OK)
pygments >=2.0 : 2.2.0 (OK)
pylint >=0.25 : 1.6.4 (OK)
qtconsole >=4.2.0: 4.3.0 (OK)
rope >=0.9.4 : 0.9.4-1 (OK)
sphinx >=0.6.6 : 1.5.6 (OK)
sympy >=0.7.3 : 1.0 (OK)

[goanpeca]

@Diego-MX could you describe the actual workflow, or how this command works?

[ccordoba12]

I think @Diego-MX wants to write something like

view(df)

in a console and make it display a Dataframe directly, instead of having to double click on the Variable Explorer.

[Diego-MX]

I have edited my question. Thank you.

[goanpeca]

@ccordoba12 what do you think? I guess we could have, spyder magics for ipython?

[ccordoba12]

We already have one called %varexp to create plots from the Variable Explorer. We could extend it to provide something like this.

[goanpeca]

We already have one called %varexp to create plots from the Variable Explorer. We could extend it to provide something like this.

Sounds good. I think a magic is better suited for this right?

[ccordoba12]

Yep, it's the right concept to use.

[tjhallum]

I would very much like a feature like the one described above by Diego-MX. Was a feature like this ever implemented?

[ccordoba12]

No, sorry. We plan to implement it in Spyder 5, to be released in 2020.

[tjhallum]

@ccordoba12 Sounds good, thank you for the expedient reply!

[akdor1154]

another use case: would be make it easier to view a subset of a large dataframe without having to create a temporary variable.
e.g. explore(df['TheColumnICareAbout', 'AnotherColumn'])

[ccordoba12]

@akdor1154, that won't be possible.

@impact27, I think this should be very easy to implement by using the new comms architecture and adding a %view magic to our kernel.

[akdor1154]

@ccordoba12, is there an architecture where that would be possible? e.g. IPython's display() achieves this workflow in notebooks, and it's really handy.

[ccordoba12]

Then you're welcome to dive in our architecture and implement it, if you think it's possible.

[impact27]

I think the best would be to have a spyder namespace for spyder magic to cleanly separate from ipython magic. something like %spyder display df or %spyder.display df.

What I would see is:
In spyder:

• in spyder/plugins/variableexplorer, define a new function to display a variable that is not in the variable explorer (for @akdor1154 suggestion) Or that finds the index of the object.
• in spyder/plugins/ipythonconsole/widgets/shell.py: register a new call_handler: eg. 'display_variable': self.display_variable that will send the received object (or object name if only the index is needed) to the variable explorer.

In spyder-kernels:

• Register the magic with ipython MagicsManager.register_function (in SpyderKernel I think this is accessed by self.shell.magics_manager.register_function)
• Send the data to be displayed to the frontend with a comm message to display_variable

[papergr]

another use case: would be make it easier to view a subset of a large dataframe without having to create a temporary variable.
e.g. explore(df['TheColumnICareAbout', 'AnotherColumn'])

As a previously RStudio user transiting to Spyder this would be extremely useful; keyboard use only plus record-wise filtering saves you in cases of significantly sized variables where variable explorer window cannot open.