Merging rows

[Shahin-rmz]

Hello
I have following table.

col1	col2	col3	col4	col5
Key1
	value1
		value2
			value3
				value4
key2

I want to merge rows in the way it shows normal table

col1	col2	col3	col4	col5
Key1	value1	value2	value3	value4
key2	value6	value7	value8	value9

how is that possible with Visidata?
Thanks

[saulpw]

Use gs to select all rows, then f to fill the nulls in each column with the row above it.Then gu to unselect all rows, select all the empty/null values (g|^$ should work), and gd to delete them. What's left should be the full rows that you want.

Note that you may have to set options.null_value="" so that VisiData considers empty strings null values; or you can Shift+M to melt the sheet, go to an empty cell and press , to select all empty cells, then gzd to set them to null values).

[Shahin-rmz]

Thanks for your response.
As you said, I melted the sheet SHIFT + m and set the empty cells to null with , .
Here is how it looks like:

then gs and f

and this is how the results looks like.

However, when I select null values with g|^$ and then gd, I lost first 1000 rows.

Would you please take a look at raw data
and also the data after doing your guide and tell me what I have done wrong?

These are multiple choice questions and their answers of an exam and some questions (first column) are multiple lines (multiple rows). This makes me a bit confused.
Maybe it would be better to merge the questions into one row and then do your manual.
bit confusing.
Thanks.

[saulpw]

The g|^$ command selects all rows with at least one cell with empty contents. So I guess your first 1000 rows have that extra column with null values. You can either remove those columns, or find a different way to select the partially-filled rows.

[Shahin-rmz]

@saulpw I was able to make the project done, with Pandas, (in Jupyter).

However I was thinking with myself, it would be great if we could have interactive DataFrame of pandas which we can work in VisiData.
I think it was mentioned before about being able to write python code inside VisiData, but this one is just pure live DataFrame.
I think it brings VisiData to another level.
Best

[saulpw]

Hi @Shahin-rmz, try using vd.view_pandas(df), which shows the pandas dataframe within VisiData.

[Shahin-rmz]

How about having live pandas dataframe and be able to do advanced python on it?
Use the true power of VisiData and Python together
Check this out.