基于多个其他列有条件地替换数据帧列中的值 - R.

Question

My data frame looks like this

> tornado_frame
         tornado_names Level      value
1     node per cluster   low  -34.72222
2          TB per node   low  -52.08333
3  expense per cluster   low -104.16667
4             Total TB   low  -62.50000
5  revenue per cluster   low  -52.08333
6     node per cluster  high   20.83333
7          TB per node  high   41.66667
8  expense per cluster  high   52.08333
9             Total TB  high  145.83333
10 revenue per cluster  high  156.25000

I want the table to transform into this

> tornado_frame
         tornado_names Level      value
1     node per cluster   low   34.72222
2          TB per node   low   52.08333
3  expense per cluster   low  104.16667
4             Total TB   low  -62.50000
5  revenue per cluster   low  -52.08333
6     node per cluster  high  -20.83333
7          TB per node  high  -41.66667
8  expense per cluster  high  -52.08333
9             Total TB  high  145.83333
10 revenue per cluster  high  156.25000

Where the negative sign in "value" changes if its absolute value is greater than that of the "high" Level column and of the same tornado_name column.

I tried a few nested if's but that got messy for me. Any help would be appreciated!

Here is my data:

> dput(tornado_frame)
structure(list(tornado_names = structure(c(2L, 4L, 1L, 5L, 3L, 
2L, 4L, 1L, 5L, 3L), .Label = c("expense per cluster", "node per cluster", 
"revenue per cluster", "TB per node", "Total TB"), class = "factor"), 
    Level = structure(c(2L, 2L, 2L, 2L, 2L, 1L, 1L, 1L, 1L, 1L
    ), .Label = c("high", "low"), class = "factor"), value = c(34.72222, 
    52.08333, 104.16667, -62.5, -52.08333, -20.83333, -41.66667, 
    -52.08333, 145.83333, 156.25)), .Names = c("tornado_names", 
"Level", "value"), class = "data.frame", row.names = c(NA, -10L
))

Answer 1

Here's a possible data.table solution

library(data.table)
setDT(df)[, value := if(diff(abs(value)) < 0) value * -1,
                                            by = tornado_names]
df
#           tornado_names Level     value
#  1:    node per cluster   low  34.72222
#  2:         TB per node   low  52.08333
#  3: expense per cluster   low 104.16667
#  4:            Total TB   low -62.50000
#  5: revenue per cluster   low -52.08333
#  6:    node per cluster  high -20.83333
#  7:         TB per node  high -41.66667
#  8: expense per cluster  high -52.08333
#  9:            Total TB  high 145.83333
# 10: revenue per cluster  high 156.25000

This will check your condition per tornado_names and only change the sign for the values within the groups where the condition is satisfied.