我需要从分组列表中读取一堆文件，然后根据组进行合并（来自同一组的文件将具有相同的列，因此可以使用bind_rows()进行缩小。
当我通过purrr::map()函数移动时，由于似乎不断收到关于不能在原子向量上使用$的警告，我似乎无法掌握数据的变化方式。
我要做的第一件事是按组划分，以便获得要在每个组中读取的文件列表的列表。然后，我使用map浏览该列表中的每个项目，并使用第二张map浏览每个子列表中的行以读取文件。但是，在该级别上发生了一些事情，不再像对待顶级团队那样处理数据了。
（缺少能够调试和查看map函数内部环境的问题，这实际上是理解机制的问题。）
require(tidyverse)
#> Loading required package: tidyverse
x <- structure(list(survey = c("adm2014", "adm2015", "adm2016", "eap2008",
"eap2009", "eap2011", "eap2012", "eap2013", "eap2014", "eap2015",
"eap2016", "ef2008a", "ef2008b", "ef2008c", "ef2008cp", "ef2008d",
"ef2009a", "ef2009b", "ef2009c", "ef2009d", "ef2010a", "ef2010b",
"ef2010c", "ef2010cp", "ef2010d", "ef2011a", "ef2011b", "ef2011c",
"ef2011d", "ef2012a", "ef2012b", "ef2012c", "ef2012cp", "ef2012d",
"ef2013a", "ef2013b", "ef2013c", "ef2013d", "ef2014a", "ef2014b",
"ef2014c", "ef2014cp", "ef2014d", "ef2015a", "ef2015b", "ef2015c",
"ef2015d", "ef2016a", "ef2016b", "ef2016c", "ef2016cp", "ef2016d",
"efest2008", "efest2009", "effy2008", "effy2009", "effy2010",
"effy2011", "effy2012", "effy2013", "effy2014", "effy2015", "effy2016",
"effy2017", "efia2008", "efia2009", "efia2011", "efia2012", "efia2013",
"efia2014", "efia2015", "efia2016", "efia2017", "f0708_f1a",
"f0708_f2", "f0708_f3", "f0809_f1a", "f0809_f2", "f0809_f3",
"f0910_f1a", "f0910_f2", "f0910_f3", "f1011_f1a", "f1011_f2",
"f1011_f3", "f1112_f1a", "f1112_f2", "f1112_f3", "f1213_f1a",
"f1213_f2", "f1213_f3", "f1314_f1a", "f1314_f2", "f1314_f3",
"f1415_f1a", "f1415_f2", "f1415_f3", "f1516_f1a", "f1516_f2",
"f1516_f3", "gr2008", "gr2008_l2", "gr2009", "gr2009_l2", "gr200_08",
"gr200_09", "gr200_10", "gr200_11", "gr200_12", "gr200_13", "gr200_14",
"gr200_15", "gr200_16", "gr2010", "gr2010_l2", "gr2011", "gr2011_l2",
"gr2012", "gr2012_l2", "gr2013", "gr2013_l2", "gr2014", "gr2014_l2",
"gr2015", "gr2015_l2", "gr2016", "gr2016_l2", "hd2008", "hd2009",
"hd2010", "hd2011", "hd2012", "hd2013", "hd2014", "hd2015", "hd2017",
"ic2008", "ic2008_ay", "ic2008_py", "ic2009", "ic2009_ay", "ic2009_py",
"ic2010", "ic2010_ay", "ic2010_py", "ic2011", "ic2011_ay", "ic2011_py",
"ic2012", "ic2012_ay", "ic2012_py", "ic2013", "ic2013_ay", "ic2013_py",
"ic2014", "ic2014_ay", "ic2014_py", "ic2015", "ic2015_ay", "ic2015_py",
"ic2016", "ic2016_ay", "ic2016_py", "ic2017", "ic2017_ay", "ic2017_py",
"s2008_abd", "s2008_cn", "s2008_f", "s2008_g", "s2009_abd", "s2009_cn",
"s2009_f", "s2009_g", "s2010_abd", "s2010_cn", "s2010_f", "s2010_g",
"s2011_abd", "s2011_cn", "s2011_f", "s2011_g", "sal2008_a", "sal2008_a_lt9",
"sal2008_b", "sal2008_faculty", "sal2009_a", "sal2009_a_lt9",
"sal2009_b", "sal2009_faculty", "sal2010_a", "sal2010_a_lt9",
"sal2010_b", "sal2010_faculty", "sal2011_a", "sal2011_a_lt9",
"sal2011_faculty"), survgroup = c("adm", "adm", "adm", "eap",
"eap", "eap", "eap", "eap", "eap", "eap", "eap", "efa", "efb",
"efc", "efcp", "efd", "efa", "efb", "efc", "efd", "efa", "efb",
"efc", "efcp", "efd", "efa", "efb", "efc", "efd", "efa", "efb",
"efc", "efcp", "efd", "efa", "efb", "efc", "efd", "efa", "efb",
"efc", "efcp", "efd", "efa", "efb", "efc", "efd", "efa", "efb",
"efc", "efcp", "efd", "efest", "efest", "effy", "effy", "effy",
"effy", "effy", "effy", "effy", "effy", "effy", "effy", "efia",
"efia", "efia", "efia", "efia", "efia", "efia", "efia", "efia",
"f_f1a", "f_f2", "f_f3", "f_f1a", "f_f2", "f_f3", "f_f1a", "f_f2",
"f_f3", "f_f1a", "f_f2", "f_f3", "f_f1a", "f_f2", "f_f3", "f_f1a",
"f_f2", "f_f3", "f_f1a", "f_f2", "f_f3", "f_f1a", "f_f2", "f_f3",
"f_f1a", "f_f2", "f_f3", "gr", "gr_l2", "gr", "gr_l2", "gr_08",
"gr_09", "gr_10", "gr_11", "gr_12", "gr_13", "gr_14", "gr_15",
"gr_16", "gr", "gr_l2", "gr", "gr_l2", "gr", "gr_l2", "gr", "gr_l2",
"gr", "gr_l2", "gr", "gr_l2", "gr", "gr_l2", "hd", "hd", "hd",
"hd", "hd", "hd", "hd", "hd", "hd", "ic", "ic_ay", "ic_py", "ic",
"ic_ay", "ic_py", "ic", "ic_ay", "ic_py", "ic", "ic_ay", "ic_py",
"ic", "ic_ay", "ic_py", "ic", "ic_ay", "ic_py", "ic", "ic_ay",
"ic_py", "ic", "ic_ay", "ic_py", "ic", "ic_ay", "ic_py", "ic",
"ic_ay", "ic_py", "s_abd", "s_cn", "s_f", "s_g", "s_abd", "s_cn",
"s_f", "s_g", "s_abd", "s_cn", "s_f", "s_g", "s_abd", "s_cn",
"s_f", "s_g", "sal_a", "sal_a_lt9", "sal_b", "sal_faculty", "sal_a",
"sal_a_lt9", "sal_b", "sal_faculty", "sal_a", "sal_a_lt9", "sal_b",
"sal_faculty", "sal_a", "sal_a_lt9", "sal_faculty")), class = c("tbl_df",
"tbl", "data.frame"), row.names = c(NA, -197L))
x %>%
split(.$survgroup) %>%
map(function(currentgroup) {
#currentgroup should now be a tibble of each group.
currentgroup %>%
map(function(singlesurvey) { #singlesurvey should be each row in the group
x <- read_csv(path_expand(paste0("~data/IPEDS/API Pulls/datadownloaded/", singlesurvey$survey, ".csv")))
}) %>% bind_rows()
})
#> Error in path_expand(paste0("~data/IPEDS/API Pulls/datadownloaded/", singlesurvey$survey, : could not find function "path_expand"
由reprex package（v0.2.1）于2018-11-12创建
答案 0 :(得分：2)
问题在于我们需要遍历列中的各个文件，而不是遍历数据集中的列。在OP的帖子中，第二个map在data.frame中循环显示一列。在这里，基本单位是带有一列的data.frame。如果将列提取为vector，则该单位将变为vector，并循环遍历vector
x %>%
split(.$survgroup) %>%
map(~ .x %>%
pull(survey) %>%
map(~ .x %>%
paste0("~data/IPEDS/API Pulls/datadownloaded/", ., '.csv') %>%
path.expand %>%
read_csv)))
答案 1 :(得分：1)
另一种解决方案是使用列表列将数据帧读入列，然后进行拆分。
x %>%
mutate(data = map(survey, ~ read_csv(path.expand(paste0("~data/IPEDS/API Pulls/datadownloaded/", .x, ".csv"))))) %>%
unnest() %>%
split(.$survgroup)