Question

What is the difference between the following git commands:

git  checkout branch
git  checkout branch .
git  checkout  .  #<-- used at the branch

Why when I checkout different branches into different folders with first one I missed some files.
But when I am using the second command, everything is ok?

Answer 1

git checkout(1) does very different things whether given path specifier or not.

1. With branch specifier only (git checkout branch) it will switch current working directory to specified branch, keeping local changes if possible and failing otherwise. If you already are on branch, it will do nothing at all. It only modifies files in the working directory that differ between HEAD and branch and fails if any of them has local modifications (indexed or not).
2. With path specifier it will overwrite the all matching files (all files match .) with specified content:
   1. With path specifier only (git checkout .) it writes content from index. That is, it undoes unstaged local modification. To undo staged modifications, use git reset with path specifier.
   2. With both branch and path specifiers (git checkout branch .) it writes content in specified revision. It will not modify where HEAD points, so if branch is different from HEAD, there will be unstaged changes afterwards.

Note, that the man page distinguishes additional cases for use of the -b/–branch option and -p/–patch option, but those are mostly straightforward extensions of the above cases.