R 在 R 中展平不同长度的嵌套列表(Google 地理编码 API 输出)

Question

我一直在使用 Google 的地理编码 API 对地址列表进行地理编码。它在嵌套列表中返回结果。列表中的元素可能会有所不同，有时会出现部分匹配，从而导致嵌套在最高级别的多个嵌套列表。到目前为止，我已将每个 GoogleResult 保存到单个数据框单元格中。

这是我的数据框的示例:

    df <- structure(list(address = structure(c(3L, 1L, 2L), .Label = c("115 Civic Parade, Altona VIC 3018", 
"Civic Parade, Altona VIC 3018", "EAST LA CLARKEFIELD 3430"), class = "factor"), 
    GoogleResult = list(list(list(access_points = list(), address_components = list(
        list(long_name = "Los Angeles", short_name = "Los Angeles", 
            types = list("locality", "political")), list(long_name = "Los Angeles County", 
            short_name = "Los Angeles County", types = list("administrative_area_level_2", 
                "political")), list(long_name = "California", 
            short_name = "CA", types = list("administrative_area_level_1", 
                "political")), list(long_name = "United States", 
            short_name = "US", types = list("country", "political"))), 
        formatted_address = "Los Angeles, CA, USA", geometry = list(
            bounds = list(northeast = list(lat = 34.3373061, 
                lng = -118.1552891), southwest = list(lat = 33.7036519, 
                lng = -118.6681759)), location = list(lat = 34.0522342, 
                lng = -118.2436849), location_type = "APPROXIMATE", 
            viewport = list(northeast = list(lat = 34.3373061, 
                lng = -118.1552891), southwest = list(lat = 33.7036519, 
                lng = -118.6681759))), partial_match = TRUE, 
        place_id = "ChIJE9on3F3HwoAR9AhGJW_fL-I", types = list(
            "locality", "political")), list(access_points = list(), 
        address_components = list(list(long_name = "3430", short_name = "3430", 
            types = list("postal_code")), list(long_name = "Clarkefield", 
            short_name = "Clarkefield", types = list("locality", 
                "political")), list(long_name = "Victoria", short_name = "VIC", 
            types = list("administrative_area_level_1", "political")), 
            list(long_name = "Australia", short_name = "AU", 
                types = list("country", "political"))), formatted_address = "Clarkefield VIC 3430, Australia", 
        geometry = list(bounds = list(northeast = list(lat = -37.4364578, 
            lng = 144.8986988), southwest = list(lat = -37.5280439, 
            lng = 144.7012193)), location = list(lat = -37.497542, 
            lng = 144.8071366), location_type = "APPROXIMATE", 
            viewport = list(northeast = list(lat = -37.4364578, 
                lng = 144.8986988), southwest = list(lat = -37.5280439, 
                lng = 144.7012193))), partial_match = TRUE, place_id = "ChIJS3IdP-xX1moRkD8uRnhWBBw", 
        types = list("postal_code"))), list(list(access_points = list(), 
        address_components = list(list(long_name = "115", short_name = "115", 
            types = list("street_number")), list(long_name = "Civic Parade", 
            short_name = "Civic Parade", types = list("route")), 
            list(long_name = "Altona", short_name = "Altona", 
                types = list("locality", "political")), list(
                long_name = "Hobsons Bay City", short_name = "Hobsons Bay", 
                types = list("administrative_area_level_2", "political")), 
            list(long_name = "Victoria", short_name = "VIC", 
                types = list("administrative_area_level_1", "political")), 
            list(long_name = "Australia", short_name = "AU", 
                types = list("country", "political")), list(long_name = "3018", 
                short_name = "3018", types = list("postal_code"))), 
        formatted_address = "115 Civic Parade, Altona VIC 3018, Australia", 
        geometry = list(bounds = list(northeast = list(lat = -37.8633208, 
            lng = 144.8316509), southwest = list(lat = -37.86409, 
            lng = 144.8303929)), location = list(lat = -37.863727, 
            lng = 144.8310159), location_type = "ROOFTOP", viewport = list(
            northeast = list(lat = -37.8623564197085, lng = 144.832370880292), 
            southwest = list(lat = -37.8650543802915, lng = 144.829672919709))), 
        place_id = "ChIJBXz75NRj1moRpVRt21nooQw", types = list(
            "premise"))), list(list(access_points = list(), address_components = list(
        list(long_name = "Civic Parade", short_name = "Civic Parade", 
            types = list("route")), list(long_name = "Altona", 
            short_name = "Altona", types = list("locality", "political")), 
        list(long_name = "Hobsons Bay City", short_name = "Hobsons Bay", 
            types = list("administrative_area_level_2", "political")), 
        list(long_name = "Victoria", short_name = "VIC", types = list(
            "administrative_area_level_1", "political")), list(
            long_name = "Australia", short_name = "AU", types = list(
                "country", "political")), list(long_name = "3018", 
            short_name = "3018", types = list("postal_code"))), 
        formatted_address = "Civic Parade, Altona VIC 3018, Australia", 
        geometry = list(bounds = list(northeast = list(lat = -37.8626502, 
            lng = 144.8449271), southwest = list(lat = -37.8661171, 
            lng = 144.81081)), location = list(lat = -37.864412, 
            lng = 144.8303004), location_type = "GEOMETRIC_CENTER", 
            viewport = list(northeast = list(lat = -37.8626502, 
                lng = 144.8449271), southwest = list(lat = -37.8661171, 
                lng = 144.81081))), place_id = "EihDaXZpYyBQYXJhZGUsIEFsdG9uYSBWSUMgMzAxOCwgQXVzdHJhbGlhIi4qLAoUChIJtbGXUCti1moRKcxHhdx2QrYSFAoSCSEyccGdYdZqEXDajCF1VgQF", 
        types = list("route"))))), row.names = c(NA, -3L), class = "data.frame")

第一种情况有部分匹配，即两个嵌套的结果列表。

我的预期输出是:

将所有列表的所有元素作为列的数据框

所有列要适当命名

部分匹配有 >1 个结果，每个匹配可以是 1 行，或者只是用“address2”变量加宽数据框。无论哪种方式我都可以使用。

我试过这样的事情:

lapply(df$GoogleResult, data.frame, stringsAsFactors = FALSE)

但元素的长度不同......导致:

arguments imply differing number of rows: 0, 1

在部分匹配的情况下，结果可以显示为数据框中的两行，或者显示为一组额外的列，我不太在意..

任何帮助是极大的赞赏!

干杯，

吕克

Answer 1

你可以试试这样的:

df %>% 
 unnest(col = GoogleResult) %>% unnest(col = GoogleResult)%>%
  filter(lengths(GoogleResult)>0)%>%
  {map2(.$GoogleResult,.$address,
        ~cbind(address = .y,data.frame(fromJSON(toJSON(.x))))%>%unnest())}%>%
  plyr::rbind.fill()