使用R中的zoo包绘制多年数据重叠图

Question

我有以下数据框：

> head(giftByDay)
    giftDate    gift   yr mon day
1 2009-07-01  100.00 2009   7   1
2 2009-07-03  300.00 2009   7   3
3 2009-07-06  470.00 2009   7   6
4 2009-07-07 7436.66 2009   7   7
5 2009-07-09   50.00 2009   7   9
6 2009-07-11   25.00 2009   7  11

这份数据有3年的记录（财政年度截至6月30日），我想将它们都放在一个图表中，分别以三条线展示每年的表现。另外，我还想将美元放在y轴上，日期放在x轴上。

根据年份，我将其分成了3个部分：

> yr2009 <- subset(giftByDay, giftByDay$giftDate >= "2009-07-01" & giftByDay$giftDate < "2010-07-01")
> yr2010 <- subset(giftByDay, giftByDay$giftDate >= "2010-07-01" & giftByDay$giftDate < "2011-07-01")
> yr2011 <- subset(giftByDay, giftByDay$giftDate >= "2011-07-01" & giftByDay$giftDate < "2012-07-01")

下一步，我创建了动物园对象以进行绘图：

> yr2009$d2 <- format(as.Date(yr2009$giftDate), "%m-%d")
> x1 <- zoo(yr2009$gift, yr2009$d2)
> yr2010$d2 <- format(as.Date(yr2010$giftDate), "%m-%d")
> x2 <- zoo(yr2010$gift, x2.Date)
> yr2011$d2 <- format(as.Date(yr2011$giftDate), "%m-%d")
> x3 <- zoo(yr2011$gift, x3.Date)

问题出现在尝试绘图时：

> plot(x1, type="l", col=1)
Error in plot.window(...) : need finite 'xlim' values
In addition: Warning messages:
1: In xy.coords(x, y, xlabel, ylabel, log) : NAs introduced by coercion
2: In min(x) : no non-missing arguments to min; returning Inf
3: In max(x) : no non-missing arguments to max; returning -Inf
> points(x2, type="l", col=2)
Warning message:
In xy.coords(x, y) : NAs introduced by coercion
> points(x3, type="l", col=3)
Warning message:
In xy.coords(x, y) : NAs introduced by coercion

实际上对象中有一些内容：

> head(x1)
   01-01    01-04    01-05    01-08    01-11    01-13 
    15.0    125.0   1000.0   6350.0    200.0 291281.1 

但我无法弄清楚为什么它无法绘图。有任何建议吗？

编辑：

> giftByDayList <- split(as.xts(giftByDay), "years")
Error in as.POSIXlt.character(x, tz, ...) : 
  character string is not in a standard unambiguous format

> class(giftByDay$giftDate)
[1] "character"

我不确定是什么原因导致了这个错误。

编辑2：

> gbd <- zoo(giftByDay)
> gbd <- as.xts(gbd)
Error in xts(coredata(x), order.by = order.by, frequency = frequency,  : 
  order.by requires an appropriate time-based object
> giftByDayList <- split(as.xts(gbd), "years")
Error in xts(coredata(x), order.by = order.by, frequency = frequency,  : 
  order.by requires an appropriate time-based object

编辑3：

> giftByDayList <- split(xts(giftByDay[,-1],as.Date(giftByDay$giftDate)), "years")
> giftByDayList <- lapply(giftByDayList, toyear, 2011)
Error in `index<-.xts`(`*tmp*`, value = list(sec = c(0, 0, 0, 0, 0, 0,  : 
  unsupported ‘index’ index type of class ‘POSIXt’unsupported ‘index’ index type of class ‘POSIXlt’

> str(giftByDayList )
List of 3
 $ :An ‘xts’ object from 2009-07-01 to 2009-12-31 containing:
  Data: num [1:89, 1:4] 100 300 470 7437 50 ...
 - attr(*, "dimnames")=List of 2
  ..$ : NULL
  ..$ : chr [1:4] "gift" "yr" "mon" "day"
  Indexed by objects of class: [Date] TZ: 
  xts Attributes:  
 NULL
 $ :An ‘xts’ object from 2010-01-01 to 2010-12-31 containing:
  Data: num [1:213, 1:4] 15 125 1000 6350 200 ...
 - attr(*, "dimnames")=List of 2
  ..$ : NULL
  ..$ : chr [1:4] "gift" "yr" "mon" "day"
  Indexed by objects of class: [Date] TZ: 
  xts Attributes:  
 NULL
 $ :An ‘xts’ object from 2011-01-02 to 2011-10-26 containing:
  Data: num [1:189, 1:4] 1500 235 1000 18154 10 ...
 - attr(*, "dimnames")=List of 2
  ..$ : NULL
  ..$ : chr [1:4] "gift" "yr" "mon" "day"
  Indexed by objects of class: [Date] TZ: 
  xts Attributes:  
 NULL
> 

- screechOwl

Answer 1

1) data.frame解决方案。试试这个。我们设置了一些测试数据，然后在transform语句中计算财政年度并重新基准化每个日期。最后，我们按照财政年度将礼物与重新基准化的日期绘制在一起。请注意，这实际上只是两个语句，一个用于transform，一个用于plot。

## set up test data

giftDate <- seq(as.Date("2009-07-01"), length = 36, by = "month")
giftByDay <- data.frame(giftDate, gift = 1:36,
    yr = as.numeric(format(giftDate, "%Y")),
    mon = as.numeric(format(giftDate, "%m")),
    day = as.numeric(format(giftDate, "%d")))

## now that we have test data, calculate fiscalyear, rebase each date to 1999-2000
## and plot

giftByDay <- transform(giftByDay, fiscalyear = yr + (mon > 6),
    Date = as.Date(paste(2000 - (mon > 6), mon, day, sep = "-")))

library(lattice)
xyplot(gift ~ Date,  giftByDay, group = fiscalyear, type = "o", auto.key = TRUE)

编辑：

下面添加了动物园解决方案。

2) 动物园解决方案。在这种情况下，解决方案并不比只使用数据框简单，尽管如果特定形式的z可以用于其他计算，则可能具有优势：

library(zoo)
library(lattice)

giftByDay <- transform(giftByDay, fiscalyear = yr + (mon > 6))

z <- read.zoo(giftByDay[2:6], index = 2:4, split = 5, FUN = function(y, m, d)
        as.Date(paste(2000 - (m > 6), m, d, sep = "-")))
xyplot(z, screen = 1, col = 1:3, type = "o", auto.key = TRUE, ylab = "Gift")

这里是动物园对象z的样子：

> z
           2010 2011 2012
1999-07-01    1   13   25
1999-08-01    2   14   26
1999-09-01    3   15   27
1999-10-01    4   16   28
1999-11-01    5   17   29
1999-12-01    6   18   30
2000-01-01    7   19   31
2000-02-01    8   20   32
2000-03-01    9   21   33
2000-04-01   10   22   34
2000-05-01   11   23   35
2000-06-01   12   24   36

- G. Grothendieck

Answer 2

这个答案是我在R-help上对与xts /时间序列和图形相关的问题...的回答进行修改后得到的版本。

您可以通过将每个索引值转换为具有相同年份来实现此目的。下面的toyear函数可以实现此功能。

toyear <- function(x, year) {
  # get year of last obs
  xyear <- .indexyear(last(x))+1900
  # get index and convert to POSIXlt
  ind <- as.POSIXlt(index(x))
  # set index year to desired value
  ind$year <- year-1900
  index(x) <- ind
  # label column with year of last obs
  colnames(x) <- paste(colnames(x),xyear,sep=".")
  x
}

# split data into a list of xts objects by year
giftByDayList <- split(xts(giftByDay[,-1],as.Date(giftByDay$giftDate)), "years")
# convert each list element to be "2011"
giftByDayList <- lapply(giftByDayList, toyear, 2011)
# merge all list elements into one object
giftByDayByYear <- as.zoo(do.call(merge, giftByDayList))
# plot on one "screen"
plot(giftByDayByYear, screens=1, col=rainbow(ncol(giftByDayByYear)))

- Joshua Ulrich

---

我无法运行代码，请查看上面的编辑并查看错误信息。 - screechOwl

@acesnap：抱歉，我以为 giftByDay 是一个动物园对象。 - Joshua Ulrich

@acesnap：你创建了一个索引为c(1,2,3,...)的动物园对象。但是，xts要求索引必须是基于时间的对象（正如错误所示）。当你遇到错误时，请使用str(gbd)来检查对象的结构。在你的第一条评论后，我编辑了我的答案，向你展示如何将giftByDay转换为xts对象。 - Joshua Ulrich

代码对我来说可以运行。我怀疑您正在运行较旧版本的xts/zoo或R。 - Joshua Ulrich

好的，我会检查更新，非常感谢您的帮助。 - screechOwl