# Creating a new dataframe from another based on mapping between variables

**URL:** <https://discourse.julialang.org/t/creating-a-new-dataframe-from-another-based-on-mapping-between-variables/87805>\
**Category:** New to Julia\
**Created:** [September 26, 2022, 10:37am UTC](https://discourse.julialang.org/t/creating-a-new-dataframe-from-another-based-on-mapping-between-variables/87805 "2022-09-26T10:37:24Z")\
**Posts on this page:** 3\
**Page:** 1

<div class="post-metadata">

**Author:** ![Nash](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/nash/32/14482_2.png) [@Nash](https://discourse.julialang.org/u/Nash)\
**Post date:** [September 26, 2022, 10:37am UTC](https://discourse.julialang.org/t/creating-a-new-dataframe-from-another-based-on-mapping-between-variables/87805/1 "2022-09-26T10:37:24Z")

</div>

I have dataframe where one column contains information about student grades from 1 to 13. Let’s call that GradeA. Another column contains information about how many students got the particular GradeA.

Now consider that another grading system exists that is much less refined than GradeA, having only 5 possible grades. Let’s call that GradeB.

Also consider that GradeB maps perfectly onto GradeA, but in a not so obvious way, because GradeB’s 1, 2 and 3, say, may be GradeB’s 1 and 5 (for example) instead of the more intuitive 1 and 2.

My question is this: what function that inputs a particular grade from GradeB can create a new dataframe that contains all the grades that correspond to that grade on the GradeA scale, along with corresponding number of students?

**Example:**  
Suppose GradeB of 3 corresponds to GradeA of 1,2, and 5 on the GradeA scale, how might I create a dataframe from my original dataframe that only includes the grades 1,2 and 5 without having to write something like:

```julia
new_df = old_df[(old_df[!,:GradeA].== 1) .|| (old_df[!,:GradeA].== 2) .|| (old_df[!,:GradeA].== 5),:]

```

---

<div class="post-metadata">

**Author:** ![jd-foster](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/jd-foster/32/35824_2.png) [@jd-foster](https://discourse.julialang.org/u/jd-foster)\
**Post date:** [September 26, 2022, 11:17am UTC](https://discourse.julialang.org/t/creating-a-new-dataframe-from-another-based-on-mapping-between-variables/87805/2 "2022-09-26T11:17:22Z")

</div>

How about using joins?

```julia
using DataFrames

dfA = DataFrame(GradeA=1:5, Count=[1,1,2,3,5])
BtoA_map = [1 => [3], 2 => [4], 3 => [1,2,5]]

dfB = DataFrame(GradeB=[], GradeA=[])

for p in BtoA_map
    for i in p.second
        push!(dfB, (p.first, i))
    end
 end

leftjoin(dfB, dfA, on=:GradeA)

```

gives

```julia
5×3 DataFrame
 Row │ GradeB GradeA Count
     │ Any Any Int64?
─────┼────────────────────────
   1 │ 3 1 1
   2 │ 3 2 1
   3 │ 1 3 2
   4 │ 2 4 3
   5 │ 3 5 5

```

---

<div class="post-metadata">

**Author:** ![SteffenPL](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/steffenpl/32/206270_2.png) [@SteffenPL](https://discourse.julialang.org/u/SteffenPL)\
**Post date:** [September 26, 2022, 11:33am UTC](https://discourse.julialang.org/t/creating-a-new-dataframe-from-another-based-on-mapping-between-variables/87805/3 "2022-09-26T11:33:24Z")

</div>

@jd-foster was faster 🙂 Here is my attempt…

```julia
df = DataFrame( gradeA = [1,2,3,4,5], gradeA_count = [10,20,30,40,50] )

# define a function for mapping from gradeA to gradeB. Toy example:
function grade_map(a)
  if a in (1,2,5)
    return 1
  elseif a in (3,)
    return 2
  else 
    return 3
  end
end

df.gradeB = grade_map.( df.gradeA )
dfB = combine( groupby( df, :gradeB ), :gradeA_count => sum => :gradeB_count )

```