r语言 - 数据操作-过滤功能(dplyr)

Question

我正试图根据纳入和排除标准清理我的数据。索引和过滤函数都给了我32个变量的3787个观察值。但是，当我使用same()或all.equal()检查时，它们是不相同的。我遗漏了什么?我认为这与filter()有关。我花了一段时间才弄清楚我必须在filter()中显式地声明NA。

是否与"filter()只包含条件为TRUE的行;它排除了FALSE和NA值。如果您想保留缺失的值，请显式请求它们…"https://github.com/tidyverse/dplyr/issues/3196

analysis1 <- nhanes[(nhanes$Age >= 30 & nhanes$Age <= 79) &
!(nhanes$Cancer == 1 | nhanes$Cancer == 7 | nhanes$Cancer == 9) &
!(nhanes$Sex == 1 & (nhanes$WTkg > 160 | nhanes$WTkg < 50)) &
!(nhanes$Sex == 2 & (nhanes$WTkg > 150 | nhanes$WTkg < 45)) &
!(nhanes$HTcm > 190),]

nhanes1 <- nhanes %>%
filter((is.na(Age) | (Age >= 30 & Age <= 79 )) &
((is.na(Cancer) | !(Cancer == 1 | Cancer == 7 | Cancer == 9))) &
((is.na(Sex & WTkg) | !(Sex == 1 & (WTkg > 160 | WTkg < 50)))) &
((is.na(Sex & WTkg) | !(Sex == 2 & (WTkg > 150 | WTkg < 45)))) &
((is.na(HTcm) | !(HTcm > 190))))
identical(analysis1, nhanes1)
all.equal(analysis1, nhanes1)

输出:

[1] FALSE
[1] "Attributes: < Component “row.names”: Modes: character, numeric >"              
[2] "Attributes: < Component “row.names”: target is character, current is numeric >"
[3] "Component “ID”: 'is.NA' value mismatch: 0 in current 229 in target"            
[4] "Component “WBC”: 'is.NA' value mismatch: 334 in current 373 in target"         
[5] "Component “RBC”: 'is.NA' value mismatch: 334 in current 373 in target"         
[6] "Component “HGB”: 'is.NA' value mismatch: 334 in current 373 in target"         
[7] "Component “HCT”: 'is.NA' value mismatch: 334 in current 373 in target"         
[8] "Component “MCV”: 'is.NA' value mismatch: 334 in current 373 in target"         
[9] "Component “MCH”: 'is.NA' value mismatch: 334 in current 373 in target"         
[10] "Component “HbA1c”: 'is.NA' value mismatch: 340 in current 379 in target"       
[11] "Component “HDL”: 'is.NA' value mismatch: 394 in current 433 in target"         
[12] "Component “CRP”: 'is.NA' value mismatch: 408 in current 446 in target"         
[13] "Component “TCHOL”: 'is.NA' value mismatch: 394 in current 433 in target"       
[14] "Component “TRIG”: 'is.NA' value mismatch: 2141 in current 2163 in target"      
[15] "Component “LDL”: 'is.NA' value mismatch: 2161 in current 2183 in target"       
[16] "Component “Sex”: 'is.NA' value mismatch: 0 in current 229 in target"           
[17] "Component “Age”: 'is.NA' value mismatch: 0 in current 229 in target"           
[18] "Component “Race2”: 'is.NA' value mismatch: 0 in current 229 in target"         
[19] "Component “Mstatus”: 'is.NA' value mismatch: 0 in current 229 in target"       
[20] "Component “Fpoverty”: 'is.NA' value mismatch: 542 in current 698 in target"    
[21] "Component “Income”: 'is.NA' value mismatch: 390 in current 556 in target"      
[22] "Component “WTkg”: 'is.NA' value mismatch: 223 in current 229 in target"        
[23] "Component “HTcm”: 'is.NA' value mismatch: 225 in current 229 in target"        
[24] "Component “WAISTcm”: 'is.NA' value mismatch: 370 in current 376 in target"     
[25] "Component “HIPcm”: 'is.NA' value mismatch: 362 in current 368 in target"       
[26] "Component “SBP”: 'is.NA' value mismatch: 438 in current 470 in target"         
[27] "Component “DBP”: 'is.NA' value mismatch: 438 in current 470 in target"         
[28] "Component “CHF”: 'is.NA' value mismatch: 0 in current 229 in target"           
[29] "Component “CHD”: 'is.NA' value mismatch: 0 in current 229 in target"           
[30] "Component “MI”: 'is.NA' value mismatch: 0 in current 229 in target"            
[31] "Component “Stroke”: 'is.NA' value mismatch: 0 in current 229 in target"        
[32] "Component “COPD”: 'is.NA' value mismatch: 0 in current 229 in target"          
[33] "Component “Cancer”: 'is.NA' value mismatch: 0 in current 229 in target"   

数据有问题。如果我检查两个数据集中相同变量之间的相关系数，它们是不同的。

res_cor <- analysis1 %>% 
cor_mat(Income, HGB, HCT, MCV, CRP, HDL, LDL, method = "spearman")
res_cor
res_cor1 <- nhanes1 %>% 
cor_mat(Income, HGB, HCT, MCV, CRP, HDL, LDL, method = "spearman")
res_cor1

输出:

rowname
<chr>
Income  HGB    HCT   MCV     CRP     HDL      LDL
<dbl> <dbl>  <dbl>  <dbl>  <dbl>   <dbl>    <dbl>
1   Income  1.000  0.055  0.039 -0.060  -0.150  0.087   0.014
2   HGB     0.055  1.000  0.970 0.220   -0.120  -0.250  0.120
3   HCT    0.039    0.970   1.000   0.180   -0.097  -0.230  0.130
4   MCV   -0.060    0.220   0.180   1.000   -0.150  0.150   -0.022
5   CRP   -0.150    -0.120  -0.097  -0.150  1.000   -0.200  0.074
6   HDL    0.087    -0.250  -0.230  0.150   -0.200  1.000   -0.011
7   LDL    0.014    0.120   0.130   -0.022  0.074   -0.011  1.000
rowname
<chr>
Income  HGB    HCT   MCV     CRP     HDL      LDL
<dbl>  <dbl>  <dbl>  <dbl>  <dbl>   <dbl>    <dbl>
1   Income  1.000   0.052   0.035   -0.057  -0.150  0.0910  0.0130
2   HGB     0.052   1.000   0.970   0.220   -0.120  -0.2500 0.1200
3   HCT     0.035   0.970   1.000   0.180   -0.099  -0.2300 0.1200
4   MCV    -0.057   0.220   0.180   1.000   -0.150  0.1500  -0.0230
5   CRP    -0.150   -0.120  -0.099  -0.150  1.000   -0.2000 0.0760
6   HDL     0.091   -0.250  -0.230  0.150   -0.200  1.0000  -0.0078
7   LDL     0.013   0.120   0.120   -0.023  0.076   -0.0078 1.0000

Answer 1

试试这个:首先对列进行排序，然后使用identical:下面是一个mtcars数据集的示例，其中有和没有排序的数据:

或者你可以像@Gregor Thomas在这里提出的那样使用setidff比较R中的2个数据名称是否相等

mtcars1 <- mtcars %>% 
arrange(cyl)
identical(mtcars, mtcars)
identical(mtcars1, mtcars)

library(dplyr)
nrow(setdiff(mtcars, mtcars1)) == 0 & nrow(setdiff(mtcars1, mtcars)) == 0

> identical(mtcars, mtcars)
[1] TRUE
> mtcars1 <- mtcars %>% 
+   arrange(cyl)
> identical(mtcars1, mtcars)
[1] FALSE
> nrow(setdiff(mtcars, mtcars1)) == 0 & nrow(setdiff(mtcars1, mtcars)) == 0
[1] TRUE