我不熟悉R和我有点丢失了组织我的数据。我的数据集如下所示(typeof(mData)返回列表):
key_as_string key doc_count colA colB types.buckets
2017-01-30T13:33:30.000+01:00 1.485780e+12 28 0 0 type1, type2, type3, 18, 5, 5
2017-01-30T13:34:00.000+01:00 1.485780e+12 175 0 0 type2, type1, type3, type4, 138, 19, 17, 1
...
我希望随时间绘制计数,即我想要一个图表,其中x轴包含时间戳(如POSIX)和 y轴计数每种类型一种颜色(doc_count列为+1)。我发现很多答案说我必须融化数据,但所有的例子都使用了这样的简单格式(最好是数据帧):
timestamp(as POSIX) doc_count type1 type2 ... type n
2017-01-30T13:33:30.000+01:00 28 18 5 ...
2017-01-30T13:34:00.000+01:00 175 19 138 ...
...
问题在于mData[["types"]][["buckets"]][[x]][[1]]
包含asysmmetric数据,即我不知道每行的顺序和类型数量(顺序对于绘图来说并不重要)。如果是这种情况,我想将0添加到相应的列。
如何将mData
转换为与上面相似的形式并绘制它?那就是我失去的地方。
这是result[["types"]][["buckets"]]
的示例输出:
[[1]]
key doc_count
1 type1 18
2 type2 5
3 type3 5
[[2]]
key doc_count
1 type2 138
2 type1 19
3 type3 17
4 type4 1
随着时间的推移绘制总计 doc_count 可以正常工作:
dates <- as.POSIXct(mData[["key"]]/1000,origin="1970-01-01")
# returns NA. Problably because it doesn't match the seconds. How do I parse it correctly?
#dates <- as.POSIXct(mData[["key_as_string"]],format="%Y-%m-%dT%H:%M:%S+01:00 CET",tz="Europe/Paris")
mDf <- data.frame(date=dates,doc_count=mData[[3]])
ggplot(mDf,aes(date,doc_count))+geom_line(colour=sample(1:255255255, 1))+xlab("")+ylab("Events")+
scale_x_datetime(date_breaks="24 hours",date_labels = "%d.%m.%Y %H:%M:%S")+
theme(axis.title=element_text(size=24,face="bold"), axis.text.y = element_text(angle=90, hjust=0.5))+geom_area(fill=sample(1:255255255, 1))
修改
这应该重新制作我的数据样本:
structure(list(key_as_string = c("2017-01-30T13:33:30.000+01:00",
"2017-01-30T13:34:00.000+01:00", "2017-01-30T13:34:30.000+01:00",
"2017-01-30T13:35:00.000+01:00", "2017-01-30T13:35:30.000+01:00",
"2017-01-30T13:36:00.000+01:00"), key = c(1485779610000, 1485779640000,
1485779670000, 1485779700000, 1485779730000, 1485779760000),
doc_count = c(28L, 175L, 122L, 526L, 160L, 1306L), types = structure(list(
colA = c(0L, 0L, 0L, 0L, 0L, 0L
), colB = c(0L, 0L, 0L, 0L, 0L, 0L), buckets = list(
structure(list(key = c("type1", "type2", "type3"
), doc_count = c(18L, 5L, 5L)), .Names = c("key",
"doc_count"), class = "data.frame", row.names = c(NA,
3L)), structure(list(key = c("type2", "type1",
"type3", "type4"), doc_count = c(138L, 19L, 17L,
1L)), .Names = c("key", "doc_count"), class = "data.frame", row.names = c(NA,
4L)), structure(list(key = c("type2", "type1",
"type3"), doc_count = c(60L, 42L, 20L)), .Names = c("key",
"doc_count"), class = "data.frame", row.names = c(NA,
3L)), structure(list(key = c("type1", "type2",
"type3", "type4"), doc_count = c(379L, 128L,
18L, 1L)), .Names = c("key", "doc_count"), class = "data.frame", row.names = c(NA,
4L)), structure(list(key = c("type2", "type3",
"type1"), doc_count = c(87L, 61L, 12L)), .Names = c("key",
"doc_count"), class = "data.frame", row.names = c(NA,
3L)), structure(list(key = c("type1", "type2",
"type3", "type4"), doc_count = c(1139L, 146L,
20L, 1L)), .Names = c("key", "doc_count"), class = "data.frame", row.names = c(NA,
4L)))), .Names = c("colA",
"colB", "buckets"), row.names = c(NA, 6L), class = "data.frame")), .Names = c("key_as_string",
"key", "doc_count", "types"), row.names = c(NA, 6L), class = "data.frame")
答案 0 :(得分:0)
在这里,我提取粒度数据,并将其与测量日期一起放在数据框中。您将获得一个易于使用ggplot绘制的长数据格式:
library(dplyr)
library(ggplot2)
tb <- mData[['types']][['buckets']]
dt <- mData[['key']]
pdf <- do.call(rbind,lapply(seq_len(length(tb)),
function(x) tb[[x]] %>% mutate(date = dt[x])))
ggplot(pdf, aes(x = date, y=doc_count, col=key)) + geom_line()