Filter rows with positive and appropriate negative values

In R I want to filter with a group_by of column "first" the rows where there is a positive value in "second" and the appropriated negative value in "second".
So in the case below the first two rows and the last two rows.

# creating first column
first <- c(1,2,3,4,5,6,7,1,1,2)
# creating second column
second <- c(88,52,-28,-19,39,86,-31,10,-88,-52)
df<-data.frame(first, second)

Result should be

data.frame(first=c(1,2,1,2), second=c(88,52,-88,-52))

You can use %in% with intersect in ave.

df[ave(df$second, df$first, FUN=\(x) x %in% intersect(x, -x)) == 1,]
#   first second
#1      1     88
#2      2     52
#9      1    -88
#10     2    -52

First create a column representing the absolute values of second, then group_by both first and the abs column. filter for groups that have exactly two entries while one of them is positive and the other one is negative.

library(dplyr)

df %>% 
  mutate(abs = abs(second)) %>% 
  group_by(first, abs) %>% 
  filter(n() == 2 & (sum(second > 0) == 1 & sum(second < 0) == 1)) %>% 
  ungroup()

# A tibble: 4 × 3
  first second   abs
  <dbl>  <dbl> <dbl>
1     1     88    88
2     2     52    52
3     1    -88    88
4     2    -52    52

Probably this could help

df %>%
    group_by(first, abs_second = abs(second)) %>%
    filter(n() == 2, sum(second) == 0) %>%
    ungroup() %>% 
    select(-abs_second)

which gives

# A tibble: 4 × 2
  first second
  <dbl>  <dbl>
1     1     88
2     2     52
3     1    -88
4     2    -52