将时间序列数据重塑为 r 中具有多个主题和变量的面板数据

Question

我有 raw panel data in the following format and would like to reshape it into the classic panel data 格式，然后在面板数据回归中使用它。
以数字开头的列名是 excel 日期格式的时间。重塑数据后，应该有公司和时间列以及 "Variable" 列中列出的数据项的列。来自 R:

dput(head(df.example))
        structure(list(Firm = c("ABB LTD N", "Facebook", "Nestle", "ABB LTD N", 
        "Facebook", "Nestle"), Variable = c("Price", "Price", "Price", 
        "Market Value", "Market Value", "Market Value"), `32508` = c(110.67, 
        162500, 14.355, 809735, 9.647, 2223.87), `32538` = c(110.35, 
        162000, 14.355, 809735, 8.836, 2036.94), `32568` = c(115.29, 
        16925, 14.355, 809735, 10.556, 2433.36), `32598` = c(130.61, 
        19175, 14.355, 809735, 11.744, 2707.32), `32628` = c(146.34, 
        4130, 14.355, 809735, 12.975, 162000)), row.names = c(NA, -6L
        ), class = c("tbl_df", "tbl", "data.frame"))

我尝试用 reshape2::melt 和 reshape2::cast 函数来做，但没有用。我找不到类似的问题。非常感谢。

Answer 1

我不确定你的列名是否带有数字，但解决方法是使用 data.table。（考虑到 Excel 来源 ("1900-01-01")，我将离开日期，如 ?as.Date 中所述）

library(data.table)
df2 <- melt(df,id = c("Firm","Variable"), variable.name = "date")
setDT(df2)[,date := as.Date(as.numeric(paste(date)), origin = "1900-01-01"))]
df2
#returns
        Firm     Variable       date      value
 1: ABB LTD N        Price 1989-01-02    110.670
 2:  Facebook        Price 1989-01-02 162500.000
 3:    Nestle        Price 1989-01-02     14.355
 4: ABB LTD N Market Value 1989-01-02 809735.000
 5:  Facebook Market Value 1989-01-02      9.647
 6:    Nestle Market Value 1989-01-02   2223.870
 7: ABB LTD N        Price 1989-02-01    110.350
 8:  Facebook        Price 1989-02-01 162000.000
 9:    Nestle        Price 1989-02-01     14.355
10: ABB LTD N Market Value 1989-02-01 809735.000
11:  Facebook Market Value 1989-02-01      8.836
12:    Nestle Market Value 1989-02-01   2036.940
13: ABB LTD N        Price 1989-03-03    115.290
14:  Facebook        Price 1989-03-03  16925.000
15:    Nestle        Price 1989-03-03     14.355
16: ABB LTD N Market Value 1989-03-03 809735.000
17:  Facebook Market Value 1989-03-03     10.556
18:    Nestle Market Value 1989-03-03   2433.360
19: ABB LTD N        Price 1989-04-02    130.610
20:  Facebook        Price 1989-04-02  19175.000
21:    Nestle        Price 1989-04-02     14.355
22: ABB LTD N Market Value 1989-04-02 809735.000
23:  Facebook Market Value 1989-04-02     11.744
24:    Nestle Market Value 1989-04-02   2707.320
25: ABB LTD N        Price 1989-05-02    146.340
26:  Facebook        Price 1989-05-02   4130.000
27:    Nestle        Price 1989-05-02     14.355
28: ABB LTD N Market Value 1989-05-02 809735.000
29:  Facebook Market Value 1989-05-02     12.975
30:    Nestle Market Value 1989-05-02 162000.000
         Firm     Variable       date      value

将时间序列数据重塑为 r 中具有多个主题和变量的面板数据

Reshape time-series data into panel-data with multiple subjects and variables in r

r

panel-data