Show group values when printing grouped dataframe #1632

pdeffebach · 2018-12-12T21:32:35Z

In reference to #1539 this PR now prints the values of the grouping column whenever it prints a grouped dataframe.

pdeffebach · 2018-12-12T21:36:13Z

Now the printing looks like this (note the :d => 1) etc.

df = DataFrame(a = repeat(1:4, outer = 5), b = randn(20), c = randn(20) .+ 1)
g = groupby(df, :a)
julia> g
GroupedDataFrame{DataFrame} with 4 groups based on key: :d
First Group: 5 rows
:d => 1
│ Row │ a     │ b        │ c         │ d     │
│     │ Int64 │ Float64  │ Float64   │ Int64 │
├─────┼───────┼──────────┼───────────┼───────┤
│ 1   │ 1     │ 0.365603 │ -0.164928 │ 1     │
│ 2   │ 1     │ 1.78441  │ 1.67254   │ 1     │
│ 3   │ 1     │ -1.4079  │ 0.0170639 │ 1     │
│ 4   │ 1     │ 0.421552 │ 1.20265   │ 1     │
│ 5   │ 1     │ -1.12468 │ 0.783903  │ 1     │
First Group: 5 rows
:d => 4
│ Row │ a     │ b         │ c        │ d     │
│     │ Int64 │ Float64   │ Float64  │ Int64 │
├─────┼───────┼───────────┼──────────┼───────┤
│ 1   │ 4     │ 0.244776  │ 1.30593  │ 4     │
│ 2   │ 4     │ 1.49806   │ 0.43346  │ 4     │
│ 3   │ 4     │ -0.372229 │ 1.52525  │ 4     │
│ 4   │ 4     │ 0.845078  │ 1.19766  │ 4     │
│ 5   │ 4     │ -0.668419 │ 0.286902 │ 4     │

nalimilan · 2018-12-13T16:41:21Z

I'd print this on the same line as the number of rows, maybe like this: First Group (5 rows): col=value (with a comma between multiple keys).

You'll also need to update tests.

pdeffebach · 2018-12-14T23:12:19Z

Let me know if you like this behavior (though the code can change obviously) and I will add tests

nalimilan · 2018-12-13T16:39:43Z

src/groupeddataframe/show.jl

@@ -6,31 +6,55 @@ function Base.show(io::IO, gd::GroupedDataFrame;
                   rowlabel::Symbol = :Row,
                   summary::Bool = true)
    N = length(gd)
-    keys = join(':' .* string.(names(gd.parent)[gd.cols]), ", ")
+    grouped_names = names(gd.parent)[gd.cols]


Suggested change

grouped_names = names(gd.parent)[gd.cols]

keynames = names(gd.parent)[gd.cols]

nalimilan · 2018-12-15T13:59:36Z

src/groupeddataframe/show.jl

-            print(io, "\nGroup $i: $nrows $rows")
+
+            identified_groups = [':' * string(parent_names[col], " = ", 
+                                 first(gd[i][col])) for col in gd.cols]


Wrong indentation. Please also remove trailing spaces (here and elsewhere).

I couldn't find any examples of the right indentation for constructors. So I added more line breaks? Let me know the course of action.

I'm not sure there's a clear rule for comprehensions, but here first is inside string, so it should definitely not be aligned with the bracket. I'd do this:

Suggested change

first(gd[i][col])) for col in gd.cols]

identified_groups = [':' * string(parent_names[col], " = ", first(gd[i][col]))

for col in gd.cols]

nalimilan · 2018-12-15T14:00:46Z

src/groupeddataframe/show.jl

+
+            print(io, "\nGroup $i ($nrows $rows): ")            
+            join(io, identified_groups, ", ", " and ")
+
            show(io, gd[i], summary=false,


Maybe we shouldn't print the grouping columns, since they're listed above?

I think it reads well.

As you like. But just to make sure we're talking about the same thing: I was suggesting we could drop the grouping columns from the data frame we print below, since all values are equal within a given group.

nalimilan · 2018-12-20T09:29:31Z

src/groupeddataframe/show.jl

-            print(io, "\nGroup $i: $nrows $rows")
+
+            identified_groups = [':' * string(parent_names[col], " = ", 
+                                 first(gd[i][col])) for col in gd.cols]


I'm not sure there's a clear rule for comprehensions, but here first is inside string, so it should definitely not be aligned with the bracket. I'd do this:

Suggested change

first(gd[i][col])) for col in gd.cols]

identified_groups = [':' * string(parent_names[col], " = ", first(gd[i][col]))

for col in gd.cols]

nalimilan · 2018-12-20T09:29:42Z

src/groupeddataframe/show.jl

+
+            print(io, "\nGroup $i ($nrows $rows): ")            
+            join(io, identified_groups, ", ", " and ")
+
            show(io, gd[i], summary=false,


As you like. But just to make sure we're talking about the same thing: I was suggesting we could drop the grouping columns from the data frame we print below, since all values are equal within a given group.

nalimilan · 2018-12-20T09:31:50Z

docs/src/lib/functions.md

@@ -1,54 +0,0 @@
-```@meta


This shouldn't be removed.

pdeffebach · 2018-12-20T16:11:09Z

Sorry about deleting that file, i did it on accident and I thought my git reset ... HEAD worked.

We should show the grouping columns, so that gd[i][1] behaves according to what the user sees in printing.

pdeffebach · 2018-12-20T18:23:10Z

Okay I think this is ready to be merged.

nalimilan

Thanks!

nalimilan · 2019-01-16T13:21:17Z

src/groupeddataframe/show.jl

    keystr = length(gd.cols) > 1 ? "keys" : "key"
    groupstr = N > 1 ? "groups" : "group"
    summary && print(io, "$(typeof(gd)) with $N $groupstr based on $keystr: $keys")
    if allgroups
        for i = 1:N
            nrows = size(gd[i], 1)
            rows = nrows > 1 ? "rows" : "row"
-            print(io, "\nGroup $i: $nrows $rows")
+
+            identified_groups = [':' * string(parent_names[col], " = ", first(gd[i][col])) 


I've just realized that when the grouping column is a string or symbol, it's printed as e.g. :x = a rather than :x = "a" or :x = :a. I guess we should change this (using repr or ourshowcompact)?

It should be repr. I will make a quick PR.

Show group values when printing grouped dataframe

1b2798d

pdeffebach mentioned this pull request Dec 12, 2018

Make describe work for a grouped dataframe #1443

Open

pdeffebach added 2 commits December 12, 2018 21:51

less repeat code

ff0b463

fix indentation

84ff9ba

Put all on one line

2762f10

nalimilan reviewed Dec 15, 2018

View reviewed changes

pdeffebach added 2 commits December 19, 2018 21:52

Respond to comments

461d345

whitespace

8a6301d

nalimilan reviewed Dec 20, 2018

View reviewed changes

pdeffebach added 2 commits December 20, 2018 12:19

small fixes

3978895

Test fixes

43e8378

fix to avoid rebase

085e672

nalimilan approved these changes Dec 21, 2018

View reviewed changes

nalimilan merged commit adc1043 into JuliaData:master Dec 26, 2018

nalimilan reviewed Jan 16, 2019

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Show group values when printing grouped dataframe #1632

Show group values when printing grouped dataframe #1632

pdeffebach commented Dec 12, 2018

pdeffebach commented Dec 12, 2018

nalimilan commented Dec 13, 2018

pdeffebach commented Dec 14, 2018

nalimilan Dec 13, 2018

nalimilan Dec 15, 2018

pdeffebach Dec 20, 2018

nalimilan Dec 20, 2018

nalimilan Dec 15, 2018

pdeffebach Dec 20, 2018

nalimilan Dec 20, 2018

nalimilan Dec 20, 2018

nalimilan Dec 20, 2018

nalimilan Dec 20, 2018

pdeffebach commented Dec 20, 2018

pdeffebach commented Dec 20, 2018

nalimilan left a comment

nalimilan Jan 16, 2019

bkamins Jan 17, 2019

nalimilan Jan 17, 2019

	grouped_names = names(gd.parent)[gd.cols]
	keynames = names(gd.parent)[gd.cols]

	first(gd[i][col])) for col in gd.cols]
	identified_groups = [':' * string(parent_names[col], " = ", first(gd[i][col]))
	for col in gd.cols]

Show group values when printing grouped dataframe #1632

Show group values when printing grouped dataframe #1632

Conversation

pdeffebach commented Dec 12, 2018

pdeffebach commented Dec 12, 2018

nalimilan commented Dec 13, 2018

pdeffebach commented Dec 14, 2018

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

pdeffebach commented Dec 20, 2018

pdeffebach commented Dec 20, 2018

nalimilan left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment