从数据框创建边缘列表

Question

我想从中得到：

session location sequence weight INDIVIDUAL action
a1      texas    1        10       john       Z1
a1      texas    2        5        peter      Z2
a1      texas    3        3        ben        Z1
a1      texas    4        5        peter      Z5
a2      calif    1        25       esther     Z3
a2      calif    2        5        peggy      Z2
a2      calif    3        10       greg       Z5

对此：

INDIVIDUAL1 INDIVIDUAL2 weight
john        peter       10
john        ben         10
peter       john        5
peter       ben         5
ben         john        3
ben         peter       3
peter       john        5
peter       ben         5        

我正在探索一些选项，包括使用for循环，但我有点担心，因为我的数据集变得非常大，可能需要很长时间。任何指针非常感谢！

谢谢！

Answer 1

这是一个简单的自连接方法。我将把sequence和session专栏留给你。

library(dplyr)
df %>% select(session, weight, sequence, INDIVIDUAL) %>%
  inner_join(., select(., session, INDIVIDUAL), by = "session") %>%
  rename(INDIVIDUAL1 = INDIVIDUAL.x, INDIVIDUAL2 = INDIVIDUAL.y) %>%
  filter(INDIVIDUAL1 != INDIVIDUAL2) %>%
  unique %>%
  arrange(session, sequence)
#    session weight sequence INDIVIDUAL1 INDIVIDUAL2
# 1       a1     10        1        john       peter
# 2       a1     10        1        john         ben
# 3       a1      5        2       peter        john
# 4       a1      5        2       peter         ben
# 5       a1      3        3         ben        john
# 6       a1      3        3         ben       peter
# 7       a1      5        4       peter        john
# 8       a1      5        4       peter         ben
# 9       a2     25        1      esther       peggy
# 10      a2     25        1      esther        greg
# 11      a2      5        2       peggy      esther
# 12      a2      5        2       peggy        greg
# 13      a2     10        3        greg      esther
# 14      a2     10        3        greg       peggy

Answer 2

这应该让你开始

你的数据

df <- read.table(text="session location sequence weight INDIVIDUAL action
a1      texas    1        10       john       Z1
a1      texas    2        5        peter      Z2
a1      texas    3        3        ben        Z1
a1      texas    4        5        peter      Z5
a2      calif    1        25       esther     Z3
a2      calif    2        5        peggy      Z2
a2      calif    3        10       greg       Z5", header=TRUE, stringsAsFactors=FALSE)

library(tidyverse)
ans <- df %>%
        group_by(session, location) %>%
        nest(INDIVIDUAL, weight) %>%
        mutate(data = map(data, ~cbind(expand.grid(.x$INDIVIDUAL, .x$INDIVIDUAL), expand.grid(.x$weight, .x$weight)) %>% setNames(paste0("V", 1:4)) )) %>%
        unnest() %>%
        filter(V1 != V2) %>%
        select(-V4) %>%
        arrange(session, V1)

# A tibble: 16 x 5
   # session location     V1     V2    V3
     # <chr>    <chr>  <chr>  <chr> <int>
 # 1      a1    texas    ben   john     3
 # 2      a1    texas    ben  peter     3
 # 3      a1    texas    ben  peter     3
 # 4      a1    texas   john  peter    10
 # 5      a1    texas   john    ben    10
 # 6      a1    texas   john  peter    10
 # 7      a1    texas  peter   john     5
 # 8      a1    texas  peter   john     5
 # 9      a1    texas  peter    ben     5
# 10      a1    texas  peter    ben     5
# 11      a2    calif esther  peggy    25
# 12      a2    calif esther   greg    25
# 13      a2    calif   greg esther    10
# 14      a2    calif   greg  peggy    10
# 15      a2    calif  peggy esther     5
# 16      a2    calif  peggy   greg     5