R language -- readr package reads and writes data

---

Preface

When you try your best ,read.csv and read.table Still unable to read data , Maybe you can try readr In bag read_table,read_csv Such as function . Try it ！

I am a very single-minded person （ lazy ） People who , Although some people on the Internet say readr Reading data in the package is faster than traditional functions , I have always used the traditional method , I haven't tried readr package .
Until today, the senior brother of the laboratory told me that he r Language cannot read data , I tried for a long time , Many parameters are used , I also searched online for a long time , But it didn't work （ Maybe I'm stupid ）, I just tried to use this package , The result is , There are not too many parameters passed in , Too cattle , I cried to death .
So record the whole process and compare the speed of reading and writing data between it and traditional functions .

---

One 、 Find the problem

I won't upload the data , Maybe you won't encounter this kind of problem .. Mainly look at the error report ！
At first, I thought it might be the problem of data coding , then ：

a <- read.csv(file = 'seerv.csv',header = T,fill = T,check.names = F,fileEncoding = 'GBK')
a <- read.csv(file = 'seerv.csv',header = T,fill = T,check.names = F,fileEncoding = 'UTF-8')

About the inside header,fill,check.names,encoding,fileEncoding And other parameters I have tried （ In many cases , The change of these parameters can smoothly read the data ）, To no avail .（ Want to try and communicate further , I can send some data privately ）

Two 、 To analyze problems

In addition to changing some parameters , I also want to enjoy it read.table Source code , It turns out that the inner part may be quoted C Language , However, I only learned C Language foundation , So the way to solve the problem through the source code is broken again .
adopt read.table, No parentheses . perhaps edit(read.table) You can see the code .

This is part of the source code ：
Error message “no lines available in input” Appear suddenly , I know where there may be a problem , But I have no way to further analyze the reason .

3、 ... and 、 solve the problem

Finally, I searched the Internet ,R Language reading data , Discovery can be achieved by readr Package read / write data , Then I successfully read .
Then I , I am lazy , I really don't want to understand why the traditional method will report an error when reading that data , Let's call it a day .

library(readr)
a <- read_csv(file = 'seer.csv')

I *, In the past, it was not used check.names=F Parameter words , Traditional methods encode some special symbols as .（ spot ）, and read_csv There is no need to use anything check.names Parameters .

Four 、 Read/write speed

Last , Many people say that this package can read and write data quickly , I'll test the difference in reading and writing speed between it and traditional functions , How big is the difference .

d <- matrix(rnorm(25000000),nrow = 5000,ncol = 5000)
d <- as.data.frame(d)
t1 <- Sys.time()
write.csv(d,file = 'write.csv.csv')
Sys.time()-t1
t1 <- Sys.time()
write_csv(d,file = 'write_csv.csv')
Sys.time()-t1

t1 <- Sys.time()
a <- read.csv(file = 'write.csv.csv')
Sys.time()-t1

t1 <- Sys.time()
b <- read_csv(file = 'write_csv.csv')
Sys.time()-t1

Write data ：

55.7 second vs 1.7 second
write_csv： The speed of write.csv： Close to 30 times .

Read in the data ：

1.3 branch vs 20.3 second
read_csv： The speed of reading data is read.csv Close to 3 times .

Compare only one data set with a pair of functions , Although the result speed gap may not be accurate , but readr The reading and writing speed of the package is really fast , I can't test every data .

Here is just a simple comparison of the reading and writing speed of a pair of functions
The parameters in the function are not mentioned .

---

summary

I don't want to understand why the data can't be read , As long as you can read it , Otherwise, elder martial brother tortured me . Ha ha ha ！

Make complaints

Say a wordless thing I met today , So and so for the purchase code , The code can't run anymore , Wrong report , Send me code and data , I ran his code and found , There is no preset gene name in the gene matrix , Just change a gene to study （ Maybe the instructions are not detailed enough ）. And then there's no then , Not a word of thanks .

A little white , Welcome criticism .
If you can solve the previous read.csv The problem of , Talk to me in private , I'll give you some data .