从r中的非结构化文本文件中提取表

Posted 2023-02-19

【中文标题】从r中的非结构化文本文件中提取表

【英文标题】：Extract table from unstructured text file in r

【发布时间】：2021-11-29 08:48:35

【问题描述】：

我有一个名为 data.txt 的文本文件，其中包含以下格式的多个表格。

//
// TABLE ET_ARCMAT
// ARCID                  MATID                                    VALTO      VALFR      MINLO                  MAXLO                  MAXFL  TRVOL                  TCONA                  TCONB                  TCONC                  TCOST                  SUBCO  TPRIO                  HORTO      RECTI
// STRING                 STRING                                   INT        INT        REAL                   REAL                   STRING REAL                   REAL                   REAL                   REAL                   REAL                   STRING REAL                   INT        INT
   01KNkQkk7jAziGNzdcbHDW 051MjWuW7jkNeBn4hOHngG                   738487     738068     0.000000000000000E+00  0.000000000000000E+00  o      0.000000000000000E+00  1.000000000000000E+00  0.000000000000000E+00  0.000000000000000E+00  1.000000000000000E-02  o      0.000000000000000E+00  0          0
   01KNkQkk7jAziGNzdcbHDW Ww7Em5Fm7jk1dsFo2QZ130                   738487     738068     0.000000000000000E+00  0.000000000000000E+00  o      0.000000000000000E+00  1.000000000000000E+00  0.000000000000000E+00  0.000000000000000E+00  1.000000000000000E-02  o      0.000000000000000E+00  0          0
   01KNkQkk7jESVQys4BYhm Ww7Em5Fm7jk1dsFo2QZ130                   738487     738068     0.000000000000000E+00  0.000000000000000E+00  o      0.000000000000000E+00  1.000000000000000E+00  0.000000000000000E+00  0.000000000000000E+00  1.000000000000000E-02  o      0.000000000000000E+00  0          0
   01KNkQkk7jESWTyY4T2hm 051MjWuW7jkNeBn4hOHngG                   738487     738068     0.000000000000000E+00  0.000000000000000E+00  o      0.000000000000000E+00  1.000000000000000E+00  0.000000000000000E+00  0.000000000000000E+00  1.000000000000000E-02  o      0.000000000000000E+00  0          0

// TABLE ET_ARC
// ARCID                  FLEID                  LOCFR                  LOCTO                  TTYPE                  DURAT      RNDTR                  TCTYP                  DISCR  RLDUR
// STRING                 STRING                 STRING                 STRING                 STRING                 INT        REAL                   REAL                   STRING REAL
   01KNkQfB7jAyV0dDRB2r0 01KNkQfB7jAyV0dDRB2r0 01KNkQUr7jAkokBn7kGH7W 01KNkQUr7jAkokBn7kEn7W 0003                   0          1.000000000000000E+00  0.000000000000000E+00  o      0.000000000000000E+00
   01KNkQfB7kE5kmrh1M5ImW 01KNkQfB7kE5kmrh1M5ImW 01KNkQUr7jAkokBn7kGH7W 01KNkQUr7jAkokBn7kFn7W 0003                   0          1.000000000000000E+00  0.000000000000000E+00  o      0.000000000000000E+00
   01KNkQkk7jAIb2l2O8wQ2G 01KNkQkk7jAIb2l2O8wQ2G ULkfcYL4YP1HCP6yexG avYtOUGQB4MYMCUioDjT9W 0003                   0          1.000000000000000E+00  0.000000000000000E+00  o      0.000000000000000E+00

//
// TABLE ET_DEMCLTIM
// DEMCL      LOCID                  MATID                                    BCKTO      LAPEN                  MAXLA      NDPEN                  PENGR
// INT        STRING                 STRING                                   INT        REAL                   INT        REAL                   INT
   1          01KNkQUr7jAkokBn7kEn7W 051Meqlv7jgqwRJvGAyX1m                   69         4.500000000000000E+08  21         9.999999999000000E+09  0
   2          01KNkQUr7jAkokBn7kEn7W 051Meqlv7jgqwRJvGAyX1m                   69         4.500000000000000E+08  21         9.999999999000000E+09  0
   3          01KNkQUr7jAkokBn7kEn7W 051Meqlv7jgqwRJvGAyX1m                   69         4.500000000000000E+08  21         9.999999999000000E+09  0

我需要提取所有表格并将它们存储在一个 excel 文件中。我试图搜索代码，但找不到正确的代码。请在这方面帮助我。

【参考方案1】：

你可以试试这个代码，它对我共享的示例很有效。

#read the data line wise
tmp <- readLines('data.txt')
#Get the position of table names
table_name_index <- grep('TABLE', tmp)
#Get the name of the tab;e
table_names <- sub('.*TABLE\\s+', '', tmp[table_name_index])
#Drop datatypes STRING, INT etc and the tablename
tmp1 <- tmp[-c(table_name_index + 2, table_name_index)]
#Drop empty rows
tmp1 <- tmp1[!(tmp1 == "//" | tmp1 == "")]
#split each table separately
setNames(by(tmp1, cumsum(grepl('//', tmp1)), function(x) 
  #Remove '//' from the column names lines
  x[1] <- sub('//\\s+', '', x[1])
  #Read the data as dataframe
  read.table(text = paste0(x, collapse = '\n'), header = TRUE)
), table_names) -> list_of_tables

#Write to excel
writexl::write_xlsx(list_of_tables, 'result.xlsx')

数据以 3 个单独的工作表存储在 excel 中。