My database has 3 tables: table1, table2 and table3
I would like to do a mysqldump on this database with the following conditions:
- Dump structure for all tables
- Only dump data for table1 and table2, ignore data in table3
Currently, I do this with 2 mysqldump statements
mysqldump -u user -p -d db > db_structure.sql
mysqldump -u user -p db --ignore-table=db.table3 &g开发者_Python百科t; table1_and_table2_data.sql
Import them in the same order they were dumped (structure, then data from table1 and table2)
Is there a way to combine this into a single mysqldump command?
You can't combine them in one command but you can execute both commands at the same time and output to the same file.
mysqldump -u user -p --no-data db > structure.sql; mysqldump -u user -p db table1 table2 >> structure.sql
to avoid having to enter the password twice you can do -ppassword
(note the lack of space!). Also use --no-data
in the first command or you end up with the data as well. -d isn't needed when you're doing just one database.
Given you may want to pipe the output to another command, as I did, instead of just redirecting to a file and appending to that file in the next command, you could try (modified from the example of stask):
(mysqldump -u $1 -p$2 -d db && mysqldump -u $1 -p$2 db --ignore-table=db.table3) |\
your_command
... in my case:
(mysqldump -u $1 -p$2 -d db && mysqldump -u $1 -p$2 db --ignore-table=db.table3) |\
gzip -9 > filename.sql.gz
Enclosing the two mysqldump commands in parentheses creates a subshell whose output we pipe into gzip and then redirect that into a file.
PS: I've also been unable to combine it into one single mysqldump invocation, though.
I don't think you can do it in one command. But you definitely can merge the output to one file. Why not to wrap it in some shell script that does following:
mysqldump -u $1 -p$2 -d db > dump.sql && mysqldump -u $1 -p$2 db --ignore-table=db.table3 >> dump.sql
You will run this script with two parameters: username and password.
It's actually pretty simple, use --where clauses on the tables where you don't want data and give it an always false condition. For instance, load data on foo and gah and only schema on bar:
mysqldump -u ... -p... myDatabase foo bar --where='1=2' gah > myfile.sql
So YES you can do this on one line.
You can remove the INSERT INTO ...
part:
mysqldump \
--opt \
-u ${DB_USER} -p${DB_PASS} \
${DB_NAME} \
| grep -v 'INSERT INTO `table3`' \
| grep -v 'INSERT INTO `table4`'
精彩评论