读取R

时间:2017-12-19 18:54:54

标签: r csv output dbf

我在C:\ Deskop \ Test文件夹中有多个DBF文件。例如,我在下面列出了5个DBF文件名。

week_1_avg_199811_var86_TBL.dbf
week_1_avg_200510_var86_TBL.dbf
week_2_avg_200009_var86_TBL.dbf
week_2_avg_200309_var86_TBL.dbf
week_3_avg_200011_var86_TBL.dbf

每个文件有5列和1410行。例如,(week_1_avg_199811_var86_TBL.dbf)我将列出5行。每个DBF文件中的列1到4是相同的。

T_R          ZONE     COUNT     AREA         MEAN
T10S-R26W    1        1         0.0156       29.9412
T10S-R27W    1        1         0.0156       30.1242
T10S-R28W    1        1         0.0156       29.7399
T10S-R2E     1        1         0.0156       37.7858
T10S-R2W     1        1         0.0156       34.6999

我想输出一个读取上述(所有)DBF文件的CSV文件(Output.csv),如下所示。

Output.csv

YEAR    T_R         Sep_week_1    Oct_week_1       Nov_week_1
1998    T10S-R26W   <some_value>       <some_value>     29.9412 
1998    T10S-R27W   <some_value>       <some_value>     30.1242 
1998    T10S-R28W   <some_value>       <some_value>     29.7399 
1998    T10S-R2E    <some_value>       <some_value>     37.7858 
1998    T10S-R2W    <some_value>      <some_value>      34.6999  
2005    T10S-R26W   <some_value>      <some_value>      <some_value>
2005    T10S-R27W   <some_value>      <some_value>      <some_value>
2005    T10S-R28W   <some_value>      <some_value>      <some_value>
2005    T10S-R2E    <some_value>       <some_value>     <some_value>
2005    T10S-R2W    <some_value>       <some_value>     <some_value>

有人可以帮助我如何在R中实现这一目标?

我有以下代码读取文件夹中的所有.dbf文件,但我需要通过读取文件名来创建YEAR和每周值列。

rm(list=ls(all=TRUE))
setwd("C:/Deskop/Test")
graphics.off()

    library("foreign")
library("purrr")

readDBA <- function(file){
  df <- read.dbf(file, as.is=FALSE)
  df$fileName <- file
  return(df)
}

file.names <- list.files(pattern="*\\.dbf")

combinedData <- file.names %>% map_dfr(readDBA)
combinedData$Year <- gsub(".*_(\\d{4}).*", "\\1", combinedData$fileName)
combinedData$Month <- as.numeric(gsub(".*_\\d{4}(\\d{2}).*", "\\1", combinedData$fileName))
combinedData$Month_Abb <- month.abb[combinedData$Month]
combinedData$Week <- substr(combinedData$fileName, 6, 6)
cols <- c("Month_Abb","Week")
combinedData$Month_Week <- do.call(paste, c(combinedData[cols], sep="_wk"))
DT <- dcast(combinedData, Year + T_R + Month_Week + MEAN ~ Month_Week) 

查看&#34; Output.csv&#34;的图像文件。 enter image description here

Using Month_Week as value column: use value.var to override.

感谢任何帮助,

提前致谢。

1 个答案:

答案 0 :(得分:1)

这里的技巧是定义一个函数来读入.dba文件,并添加一个文件名标签。然后,我们将使用purrr:map_dfr来读取,然后在一个步骤中rbind()每个文件。

require(purrr)
require(foreign)

readDBA <- function(file){
  df <- read.dba(file, as.is=FALSE)
  df$fileName <- file
  return(df)
}

file.names <- list.files(pattern="*\\.dba")

combinedData <- file.names %>% map_dfr(readDBA)

然后有一点REGEX来获得年份和月份

combinedData$Year <- gsub(".*_(\\d{4}).*", "\\1", combinedData$fileName)
combinedData$Month <- gsub(".*_\\d{4}(\\d{2}).*", "\\1", combinedData$fileName)

我会让你弄清楚将周和月放在一起。

然后,您希望reshape2::dcast()采用宽格式,类似这样的内容,但除非您提供数据框的实际样本,否则无法确定。

require(reshape2)

dcast(combinedData, Year + T_R ~ MonthWeek, value.var= "MEAN")

另外,R有一个方便的month.abb向量,因此您可以使用month.abb[combinedData$Month]获取月份缩写