使用 base R 更有效地将宽数据重新格式化为长数据

Question

关于 reshape() 的其他问题帮助我重新格式化了我的数据（见下文），但我想知道是否有办法以更简单的方式（更少的代码）实现我所做的事情 using base R，没有包。我有这样的示例数据：

period store_id product price1 price2 price3 quantity1 quantity2 quantity3
201801        1     1       11      5      6       100       200       300
201802        1     1       12      6      6       100       200       300
201803        1     1       13      7      6       100       200       300
201804        1     1       14      8      6       100       200       300
201805        1     1       15      9      6       100       200       300
201806        2     2       16     10      6       100       200       300
201807        2     2       17     11      6       100       200       300
201808        2     2       18     12      6       100       200       300
201809        2     2       19     13      6       100       200       300
201810        2     2       20     14      6       100       200       300

我希望它看起来像这样（每个周期有 3 行，一列用于价格，一列用于数量）：

period store_id  product quantity type   price 
201801        1     1     100    price1    11 
201801        1     1     200    price2     5 
201801        1     1     300    price3     6 
201802        1     1     100    price1    12 
201802        1     1     200    price2     6 
201802        1     1     300    price3     6 
201803        1     1     100    price1    13 
201803        1     1     200    price2     7 
201803        1     1     300    price3     6 
201804        1     1     100    price1    14 
and so on...

我能够重新格式化我的数据，方法是首先将数据拆分为两个数据框，一个用于价格，一个用于数量，然后将它们重新合并在一起。有没有比我在下面做的更少 steps/code 更简单的方法来为我的示例使用 reshape()？

感谢任何帮助，谢谢。

CODE TO PRODUCE INITIAL DATA SET

period <- 201801:201810
df <- data.frame(period)
df$store_id <- c(1,1,1,1,1,2,2,2,2,2)
df$product <- c(1,1,1,1,1,2,2,2,2,2)
df$price1 <- 11:20
df$price2 <- 5:14
df$price3 <- df_test$price1-df_test$price2
df$quantity1 <- 100
df$quantity2 <- 200
df$quantity3 <- 300
df

MY INEFFICIENT CODE / SOLUTION

# prices
df_prices <- df[,!names(df)%in%c("quantity1","quantity2","quantity3")]

df_prices <- reshape(df_prices,               
                     direction = "long",     
                     idvar = c("period", "product", "store_id"), 
                     timevar = "type",           
                     varying = c("price1","price2","price3"),  
                     v.names = "price",     
                     times = c("price1","price2","price3"))

# quantities
df_quantities <- df[,!names(df)%in%c("price1","price2","price3")]

df_quantities <- reshape(df_quantities,               
                         direction = "long",     
                         idvar = c("period", "product", "store_id"), 
                         timevar = "type",           
                         varying = c("quantity1","quantity2","quantity3"),  
                         v.names = "quantity",     
                         times = c("quantity1","quantity2","quantity3"))

# make variables consistent to merge
df_quantities$type[df_quantities$type=="quantity1"] <- "price1"
df_quantities$type[df_quantities$type=="quantity2"] <- "price2"
df_quantities$type[df_quantities$type=="quantity3"] <- "price3"

# merge both
df <- merge(df_prices, df_quantities, by=c("period","product","store_id","type"))

Answer 1

您可以将变化指定为 4:9，即第 4 列到第 9 列。并指定 sep=''，即 time 和 1 之间没有任何内容，您有 time1.如果你有 time_1 那么 sep = '_':

reshape(df, 4:9, dir = 'long', sep="")

     period store_id product time price quantity id
1.1  201801        1       1    1    11      100  1
2.1  201802        1       1    1    12      100  2
3.1  201803        1       1    1    13      100  3
4.1  201804        1       1    1    14      100  4
5.1  201805        1       1    1    15      100  5
6.1  201806        2       2    1    16      100  6
7.1  201807        2       2    1    17      100  7
8.1  201808        2       2    1    18      100  8
9.1  201809        2       2    1    19      100  9
10.1 201810        2       2    1    20      100 10
1.2  201801        1       1    2     5      200  1
2.2  201802        1       1    2     6      200  2
3.2  201803        1       1    2     7      200  3
4.2  201804        1       1    2     8      200  4
5.2  201805        1       1    2     9      200  5
6.2  201806        2       2    2    10      200  6
7.2  201807        2       2    2    11      200  7
8.2  201808        2       2    2    12      200  8
9.2  201809        2       2    2    13      200  9
10.2 201810        2       2    2    14      200 10
1.3  201801        1       1    3     6      300  1
2.3  201802        1       1    3     6      300  2
3.3  201803        1       1    3     6      300  3
4.3  201804        1       1    3     6      300  4
5.3  201805        1       1    3     6      300  5
6.3  201806        2       2    3     6      300  6
7.3  201807        2       2    3     6      300  7
8.3  201808        2       2    3     6      300  8
9.3  201809        2       2    3     6      300  9
10.3 201810        2       2    3     6      300 10

使用 base R 更有效地将宽数据重新格式化为长数据

Using base R to reformat wide data to long data more efficiently

r

reshape