用 R 数据框中的最新数据填充缺失的日期
[英]Filling in missing dates with most recent data in data frame in R
问题描述
I have a data frame with country, date, identifier, cumulative_identifier, cumulative_country.
我有一个包含国家、日期、标识符、累积标识符、累积国家的数据框。 Country, data, and identifier are grouped.国家、数据和标识符被分组。 I, however, have countries, and identifiers with missing dates.但是，我有国家和标识符缺少日期。 These are countries that have not submitted data for these days for this identifier.这些是这些天未提交此标识符数据的国家/地区。 I would like to include these dates, but use the data from the most recent submission.我想包括这些日期，但使用最近提交的数据。
The data must be grouped by country, date, and identifier.数据必须按国家、日期和标识符分组。 For example give a data frame below.例如在下面给出一个数据框。
country date identifier cumulative_id cumulative_country
France 2021-03-20 B.1.1.7 3528 12158
France 2021-03-15 B.1.1.7 3520 12150
France 2021-03-15 B.1.2 50 12142
France 2021-03-14 B.1.2 48 12140
Morocco 2021-03-16 B.1.1.7 232 5636
Morocco 2020-03-01 B.1.1.7 220 5624
In the example above, there are many dates missing.在上面的示例中，缺少许多日期。 The dates added would use the information from the most recent submission.添加的日期将使用最近提交的信息。 So France and Morocco should look like this:所以法国和摩洛哥应该是这样的：
country date identifier cumulative_id cumulative_country
France 2021-03-20 B.1.1.7 3528 12158
France 2021-03-19 B.1.1.7 3520 12150
France 2021-03-18 B.1.1.7 3520 12150
France 2021-03-17 B.1.1.7 3520 12150
France 2021-03-16 B.1.1.7 3520 12150
France 2021-03-20 B.1.2 50 12142
France 2021-03-19 B.1.2 50 12142
France 2021-03-18 B.1.2 50 12142
France 2021-03-17 B.1.2 50 12142
France 2021-03-16 B.1.2 50 12142
France 2021-03-15 B.1.2 50 12142
France 2021-03-14 B.1.2 48 12140
France 2021-03-13 B.1.2 48 12140
Morocco 2021-03-20 B.1.1.7 232 5636
Morocco 2021-03-19 B.1.1.7 232 5636
Morocco 2021-03-18 B.1.1.7 232 5636
Morocco 2021-03-17 B.1.1.7 232 5636
Morocco 2021-03-16 B.1.1.7 232 5636
Morocco 2021-03-15 B.1.1.7 220 5624
...
Morocco 2021-03-01 B.1.1.7 220 5624
This is what I have tried with Aurèle's suggestion: The resulting date frame, however, is identical to the original, with no changes.这是我根据 Aurèle 的建议尝试过的：但是，生成的日期框架与原始日期框架相同，没有任何变化。 Again, it takes 8 minutes to complete, since there are already over 100,000 observations in the dataset.同样，这需要 8 分钟才能完成，因为数据集中已经有超过 100,000 个观察值。
horizontal$date <- as.Date(horizontal$date)
df <- df %>%
complete(nesting(country, pango_lineage), date = full_seq(date, 1)) %>%
group_by(country, pango_lineage) %>%
mutate(across(c(cum_country_pang, cum_country), zoo::na.locf, na.rm = FALSE)) %>%
filter(!is.na(cum_country_pang))
2 个解决方案
解决方案1
2 2021-04-02 14:44:55
Using tidyr complete and zoo na.locf (Last Observation Carried Forward):使用tidyr complete和zoo na.locf （最后一次观察结转）：
library(tidyr)
library(dplyr)
df %>%
complete(nesting(country, identifier), date = full_seq(date, 1)) %>%
group_by(country, identifier) %>%
mutate(across(c(cumulative_id, cumulative_country), zoo::na.locf, na.rm = FALSE)) %>%
filter(!is.na(cumulative_id))
#> # A tibble: 398 x 5
#> # Groups: country, identifier [3]
#> country identifier date cumulative_id cumulative_country
#> <chr> <chr> <date> <int> <int>
#> 1 France B.1.1.7 2021-03-15 3520 12150
#> 2 France B.1.1.7 2021-03-16 3520 12150
#> 3 France B.1.1.7 2021-03-17 3520 12150
#> 4 France B.1.1.7 2021-03-18 3520 12150
#> 5 France B.1.1.7 2021-03-19 3520 12150
#> 6 France B.1.1.7 2021-03-20 3528 12158
#> 7 France B.1.2 2021-03-14 48 12140
#> 8 France B.1.2 2021-03-15 50 12142
#> 9 France B.1.2 2021-03-16 50 12142
#> 10 France B.1.2 2021-03-17 50 12142
#> # ... with 388 more rows
Data:数据：
df <- read.table(text =
'country date identifier cumulative_id cumulative_country
France 2021-03-20 B.1.1.7 3528 12158
France 2021-03-15 B.1.1.7 3520 12150
France 2021-03-15 B.1.2 50 12142
France 2021-03-14 B.1.2 48 12140
Morocco 2021-03-16 B.1.1.7 232 5636
Morocco 2020-03-01 B.1.1.7 220 5624
', header = TRUE)
df$date <- as.Date(df$date)
解决方案2
0 2021-04-02 14:57:14
Instead of zoo::na.locf , just use tidyr::fill而不是zoo::na.locf ，只需使用tidyr::fill
library(dplyr)
library(tidyr)
df %>%
complete(nesting(country, identifier), date = full_seq(date, 1)) %>%
group_by(country, identifier) %>%
fill(c(cumulative_id, cumulative_country), .direction = "down") %>%
filter(!is.na(cumulative_id))
#> # A tibble: 398 x 5
#> # Groups: country, identifier [3]
#> country identifier date cumulative_id cumulative_country
#> <chr> <chr> <date> <int> <int>
#> 1 France B.1.1.7 2021-03-15 3520 12150
#> 2 France B.1.1.7 2021-03-16 3520 12150
#> 3 France B.1.1.7 2021-03-17 3520 12150
#> 4 France B.1.1.7 2021-03-18 3520 12150
#> 5 France B.1.1.7 2021-03-19 3520 12150
#> 6 France B.1.1.7 2021-03-20 3528 12158
#> 7 France B.1.2 2021-03-14 48 12140
#> 8 France B.1.2 2021-03-15 50 12142
#> 9 France B.1.2 2021-03-16 50 12142
#> 10 France B.1.2 2021-03-17 50 12142
#> # … with 388 more rows
Created on 2021-04-02 by the reprex package (v1.0.0)由代表 package (v1.0.0) 于 2021 年 4 月 2 日创建
声明:本站的技术帖子网页，遵循CC BY-SA 4.0协议，如果您需要转载，请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.