Есть ли способ заполнить пустые столбцы путем сопоставления терминов из одного столбца в другой, когда термины не являются точными?

Question

Я использовал rbind для объединения двух наборов данных в кадр данных 828 x 5, называемый вегетация . Вот пример моего текущего фрейма данных:

site                           year    sos    eos    vegetation
EPIC_alligatorriver.csv        2016    176    301
EPIC_alligatorriver.csv        2018    164    291
Landsat_alligatorriver.csv     2016    170    303
Pheno_alligatorriver.csv       2017    152    288
Landsat_NEON.BART42.csv        2017    115    290
Pheno_NEON.BART42.csv          2017    120    290
alligatorriver.csv             NA      NA     NA     deciduous broadleaf
NEON.BART42.csv                NA      NA     NA     mixed forest

Вот что мне нужно

site                           year    sos    eos    vegetation
EPIC_alligatorriver.csv        2016    176    301    deciduous broadleaf
EPIC_alligatorriver.csv        2018    164    291    deciduous broadleaf
Landsat_alligatorriver.csv     2016    170    303    deciduous broadleaf
Pheno_alligatorriver.csv       2017    152    288    deciduous broadleaf
Landsat_NEON.BART42.csv        2017    115    290    mixed forest
Pheno_NEON.BART42.csv          2017    120    290    mixed forest
alligatorriver.csv             NA      NA     NA     deciduous broadleaf
NEON.BART42.csv                NA      NA     NA     mixed forest

По сути, мне нужно заполнить пустой столбец растительности, сопоставляя правильный тип растительности на основе на имя сайта.

Я также могу удалить rbind и хранить все данные о растительности в совершенно другом двухстолбцовом фрейме данных

site                     vegetation
alligatorriver.csv       deciduous broadleaf
konza.csv                grassland
merbleue.csv             wetland
NEON.BART42.csv          mixed forest

, если это облегчает задачу. Всего существует 7 типов растительности и 99 участков за трехлетний период (2016, 2017, 2018). Любая помощь приветствуется!

df1 (partial): structure(list(site = c("EPIC_alligatorriver.csv", "EPIC_alligatorriver.csv", 
"EPIC_alligatorriver.csv", "EPIC_arbutuslakeinlet.csv", "EPIC_arbutuslakeinlet.csv", 
"EPIC_arbutuslakeinlet.csv", "EPIC_archboldavir.csv", "EPIC_archboldavir.csv", 
"EPIC_archboldavir.csv", "EPIC_archboldavirx.csv"), year = c(2016L, 2017L, 2018L, 2016L, 2017L, 2018L, 2016L, 
2017L, 2018L, 2016L), sos = c(117, 111, 122, 147, 145, 144, 98, 156, 
114, 98), eos = c(294, 294, 274, 276, 271, 274, 315, 295, 307, 315), vegetation = c("", "", "", "", "", "", "", "", "", "")), row.names = c("1", "2", "3", "4", "5", "6", "7", "8", "9", "10"), class = "data.frame")

df2: structure(list(vegetation = structure(c(2L, 2L, 1L, 1L, 1L, 1L, 
1L, 2L, 2L, 2L, 2L, 6L, 3L, 2L, 4L, 2L, 2L, 3L, 4L, 2L, 3L, 2L, 
1L, 3L, 3L, 2L, 3L, 2L, 2L, 3L, 6L, 6L, 4L, 1L, 4L, 1L, 4L, 2L, 
3L, 2L, 4L, 7L, 6L, 1L, 1L, 1L, 1L, 1L, 1L, 7L, 2L, 4L, 7L, 2L, 
4L, 2L, 2L, 5L, 2L, 4L, 2L, 5L, 2L, 5L, 2L, 6L, 2L, 5L, 3L, 2L, 
2L, 5L, 2L, 4L, 4L, 2L, 4L, 2L, 2L, 2L, 1L, 2L, 2L, 4L, 6L, 2L, 
1L, 4L, 2L, 1L, 3L, 2L, 2L, 4L, 3L, 2L, 3L, 2L, 2L), .Label = c("agriculture", 
"deciduous broadleaf", "evergreen needlefeaf", "grassland", "mixed forest", 
"shrub", "wetland"), class = "factor"), site = structure(1:99, .Label = c("alligatorriver", 
"arbutuslakeinlet", "archboldavir", "archboldavirx", "archboldpnot", 
"archboldpnotx", "arsmnswanlake1", "ashburnham", "bartlettir", 
"bostoncommon", "bullshoals", "burnssagebrush", "canadaOBS", 
"caryinstitute", "cperuvb", "downerwoods", "dukehw", "goodnow", 
"grandteton", "harvard", "harvardbarn", "harvardbarn2", "harvardfarmsouth", 
"harvardhemlock", "harvardhemlock2", "harvardlph", "howland1", 
"howland2", "hubbardbrook", "huyckpreserveny", "jerbajada", "jernort", 
"kansas", "kelloggcorn", "kendall", "kingmanfarm", "konza", "lacclair", 
"laclaflamme", "laurentides", "lethbridge", "lostcreek", "luckyhills", 
"mandanh5", "mandani2", "mead1", "mead2", "mead3", "meadpasture", 
"merbleue", "missouriozarks", "montebondonegrass", "montebondonepeat", 
"morganmonroe", "nationalelkrefuge", "ncssm", "NEON.BART33", 
"NEON.BART42", "NEON.DELA33", "NEON.DSNY33", "NEON.HARV33", "NEON.HARV42", 
"NEON.JERC33", "NEON.JERC42", "NEON.LENO33", "NEON.ONAQ33", "NEON.ORNL33", 
"NEON.ORNL42", "NEON.RMNP33", "NEON.SERC33", "NEON.TREE33", "NEON.TREE42", 
"NEON.UNDE33", "NEON.WOOD33", "ninemileprairie", "northattkeboroma", 
"oakville", "proctor", "queens", "readingma", "rosemountnprs", 
"russellsage", "sanford", "sevilletagrass", "sevilletashrub", 
"shalehillsczo", "southerngreatplains", "stjones", "sweetbriar", 
"sweetbriargrass", "sylvania", "tonzi", "umichbiological2", "usgseros", 
"usmpj", "uwmfieldsta", "warrenwilson", "willowcreek", "worcester"
), class = "factor")), class = "data.frame", row.names = c(NA, 
-99L))

Answer 1

Это можно сделать, используя match(), используя столбцы сайта в двух данных, используя sub(), чтобы удалить префикс из первого data.frame:

df$vegetation <- df2$vegetation[match(sub("^.*_", "", df$site), df2$site)]

df

                        site year sos eos          vegetation
1    EPIC_alligatorriver.csv 2016 176 301 deciduous broadleaf
2    EPIC_alligatorriver.csv 2018 164 291 deciduous broadleaf
3 Landsat_alligatorriver.csv 2016 170 303 deciduous broadleaf
4   Pheno_alligatorriver.csv 2017 152 288 deciduous broadleaf
5    Landsat_NEON.BART42.csv 2017 115 290        mixed forest
6      Pheno_NEON.BART42.csv 2017 120 290        mixed forest

Данные:

df <- structure(list(site = c("EPIC_alligatorriver.csv", "EPIC_alligatorriver.csv", 
                              "Landsat_alligatorriver.csv", "Pheno_alligatorriver.csv", "Landsat_NEON.BART42.csv", 
                              "Pheno_NEON.BART42.csv"), year = c(2016L, 2018L, 2016L, 2017L, 
                                                                 2017L, 2017L), sos = c(176L, 164L, 170L, 152L, 115L, 120L), eos = c(301L, 
                                                                                                                                     291L, 303L, 288L, 290L, 290L), vegetation = c("deciduous broadleaf", 
                                                                                                                                                                                   "deciduous broadleaf", "deciduous broadleaf", "deciduous broadleaf", 
                                                                                                                                                                                   "mixed forest", "mixed forest")), row.names = c(NA, -6L), class = "data.frame")

df2 <- structure(list(site = c("alligatorriver.csv", "konza.csv", "merbleue.csv", 
                               "NEON.BART42.csv"), vegetation = c("deciduous broadleaf", "grassland", 
                                                                  "wetland", "mixed forest")), class = "data.frame", row.names = c(NA, 
                                                                                                                                   -4L))