awk redirecting output in AWK: override or not?

print items > output-file

This redirection prints the items into the output file named output-file. The file name output-file can be any expression. Its value is changed to a string and then used as a file name (see Expressions).When this type of redirection is used, the output-file is erased before the first output is written to it. Subsequent writes to the same output-file do not erase output-file, but append to it. (This is different from how you use redirections in shell scripts.) If output-file does not exist, it is created. For example, here is how an awk program can write a list of BBS names to one file named name-list, and a list of phone numbers to another file named phone-list:

          $ awk '{ print $2 > "phone-list"
          >        print $1 > "name-list" }' BBS-list
          $ cat phone-list
          -| 555-5553
          -| 555-3412
          ...
          $ cat name-list
          -| aardvark
          -| alpo-net
          ...

Each output file contains one name or number per line.

print items >> output-file

This redirection prints the items into the pre-existing output file named output-file. The difference between this and the single-‘>’ redirection is that the old contents (if any) of output-file are not erased. Instead, the awk output is appended to the file. If output-file does not exist, then it is created.

Here is the example code:

cat > test.txt

a

b

c

d

awk 'BEGIN{print "e1" >> "test.txt"; print "e2" >> "test.txt"; print "e3" >> "test.txt";}'

$ cat test.txt 

a

b

c

d

e1

e2

e3

awk 'BEGIN{print "e1" > "test.txt"; print "e2" > "test.txt"; print "e3" > "test.txt";}'

$ cat test.txt 

e1

e2

e3

But I just don't understand why the test.txt is gone in code below:

awk 'BEGIN{print "e1" > "test.txt"; system("rm test.txt");print "e2" > "test.txt"; print "e3" > "test.txt";}'