Linux上删除大量文件几种方式对比

linux上删除大量小文件删除测试：删除500000个小文件测试

生成文件：

$ for i in $(seq 500000);do echo 'text' >hello/$i.txt;done

$ du -sh hello

2.0G	hello

1. rm删除：因为文件数量太多，rm无法删除（报错）

$ time rm -f *

-bash: /usr/bin/rm: Argument list too long

real	0m2.676s

user	0m2.340s

sys	0m0.272s

2. find查找删除：-exec

-exec: 该参数是将查找的结果文件名，逐个传递给后面的命令执行，如果文件较多，会导致效率较低。

xargs: 该命令是将查找的结果一次性传给后面的命令执行，效率比较高。但是不能操作文件名有特殊字符的文件。

$ time find . -type f -exec rm {} \;

real	7m16.154s

user	0m51.171s

sys	5m51.214s

3. find查找删除：xargs

$ time find . -type f|xargs rm -rf

real	0m17.860s

user	0m1.021s

sys	0m8.838s

4. find调用-delete删除

$ time find . -type f -delete

real	0m16.953s

user	0m0.535s

sys	0m8.866s

$ time find . -type f -delete

real	0m16.221s

user	0m0.578s

sys	0m8.098s

5. ls和xargs组合删除

$ time ls|xargs -L 10000 rm -f

real	0m13.764s

user	0m2.856s

sys	0m8.630s

$ time ls|xargs -L 5000 rm -f

real	0m13.879s

user	0m3.053s

sys	0m8.622s

6. rsync delete删除

--delete delete extraneous files from dest dirs

--delete-before receiver deletes before xfer, not during

--delete-during receiver deletes during the transfer

--delete-after receiver deletes after transfer, not during

$ mkdir blank # 建立空文件夹

$ time rsync -a --delete blank/ hello/

real	0m15.293s

user	0m0.590s

sys	0m12.397s

$ time rsync -a --delete blank/ hello/

real	0m14.833s

user	0m0.590s

sys	0m11.977s

$ time rsync -a --delete-before blank/ hello/

real	0m14.884s

user	0m0.602s

sys	0m12.147s

$ time rsync -a --delete-during blank/ hello/

real	0m15.468s

user	0m0.693s

sys	0m12.441s

$ time rsync -a --delete-after blank/ hello/

real	0m15.135s

user	0m0.600s

sys	0m12.253s

总结：

命令	时间	备注
rm	无法删除	...
find + -exec	7m16s	耗费时间太长
find + xargs	17s	时间较短
find + delete	16s	时间较短
ls + xargs	13s	时间短
rsync + delete	15s	时间短