Using mysqldump to format one insert per line?

问题

This has been asked a few times but I cannot find a resolution to my problem. Basically when using mysqldump, which is the built in tool for the MySQL Workbench administration tool, when I dump a database using extended inserts, I get massive long lines of data. I understand why it does this, as it speeds inserts by inserting the data as one command (especially on InnoDB), but the formatting makes it REALLY difficult to actually look at the data in a dump file, or compare two files with a diff tool if you are storing them in version control etc. In my case I am storing them in version control as we use the dump files to keep track of our integration test database.

Now I know I can turn off extended inserts, so I will get one insert per line, which works, but any time you do a restore with the dump file it will be slower.

My core problem is that in the OLD tool we used to use (MySQL Administrator) when I dump a file, it does basically the same thing but it FORMATS that INSERT statement to put one insert per line, while still doing bulk inserts. So instead of this:

INSERT INTO `coupon_gv_customer` (`customer_id`,`amount`) VALUES (887,'0.0000'),191607,'1.0300');

you get this:

INSERT INTO `coupon_gv_customer` (`customer_id`,`amount`) VALUES 
 (887,'0.0000'),
 (191607,'1.0300');

No matter what options I try, there does not seem to be any way of being able to get a dump like this, which is really the best of both worlds. Yes, it take a little more space, but in situations where you need a human to read the files, it makes it MUCH more useful.

Am I missing something and there is a way to do this with MySQLDump, or have we all gone backwards and this feature in the old (now deprecated) MySQL Administrator tool is no longer available?

回答1:

With the default mysqldump format, each record dumped will generate an individual INSERT command in the dump file (i.e., the sql file), each on its own line. This is perfect for source control (e.g., svn, git, etc.) as it makes the diff and delta resolution much finer, and ultimately results in a more efficient source control process. However, for significantly sized tables, executing all those INSERT queries can potentially make restoration from the sql file prohibitively slow.

Using the --extended-insert option fixes the multiple INSERT problem by wrapping all the records into a single INSERT command on a single line in the dumped sql file. However, the source control process becomes very inefficient. The entire table contents is represented on a single line in the sql file, and if a single character changes anywhere in that table, source control will flag the entire line (i.e., the entire table) as the delta between versions. And, for large tables, this negates many of the benefits of using a formal source control system.

So ideally, for efficient database restoration, in the sql file, we want each table to be represented by a single INSERT. For an efficient source control process, in the sql file, we want each record in that INSERT command to reside on its own line.

My solution to this is the following back-up script:

#!/bin/bash

cd my_git_directory/

ARGS="--host=myhostname --user=myusername --password=mypassword --opt --skip-dump-date"
/usr/bin/mysqldump $ARGS --database mydatabase | sed 's$VALUES ($VALUES\n($g' | sed 's$),($),\n($g' > mydatabase.sql

git fetch origin master
git merge origin/master
git add mydatabase.sql
git commit -m "Daily backup."
git push origin master

The result is a sql file INSERT command format that looks like:

INSERT INTO `mytable` VALUES
(r1c1value, r1c2value, r1c3value),
(r2c1value, r2c2value, r2c3value),
(r3c1value, r3c2value, r3c3value);

Some notes:

password on the command line ... I know, not secure, different discussion.
--opt: Among other things, turns on the --extended-insert option (i.e., one INSERT per table).
--skip-dump-date: mysqldump normally puts a date/time stamp in the sql file when created. This can become annoying in source control when the only delta between versions is that date/time stamp. The OS and source control system will date/time stamp the file and version. Its not really needed in the sql file.
The git commands are not central to the fundamental question (formatting the sql file), but shows how I get my sql file back into source control, something similar can be done with svn. When combining this sql file format with your source control of choice, you will find that when your users update their working copies, they only need to move the deltas (i.e., changed records) across the internet, and they can take advantage of diff utilities to easily see what records in the database have changed.
If you're dumping a database that resides on a remote server, if possible, run this script on that server to avoid pushing the entire contents of the database across the network with each dump.
If possible, establish a working source control repository for your sql files on the same server you are running this script from; check them into the repository from there. This will also help prevent having to push the entire database across the network with every dump.

回答2:

Try use the following option: --skip-extended-insert

It worked for me.

回答3:

As others have said using sed to replace "),(" is not safe as this can appear as content in the database. There is a way to do this however: if your database name is my_database then run the following:

$ mysqldump -u my_db_user -p -h 127.0.0.1 --skip-extended-insert my_database > my_database.sql
$ sed ':a;N;$!ba;s/)\;\nINSERT INTO `[A-Za-z0-9$_]*` VALUES /),\n/g' my_database.sql > my_database2.sql

you can also use "sed -i" to replace in-line.

Here is what this code is doing:

--skip-extended-insert will create one INSERT INTO for every row you have.
Now we use sed to clean up the data. Note that regular search/replace with sed applies for single line so we cannot detect the "\n" character as sed works one line at a time. That is why we put ":a;N;$!ba;" which basically tells sed to search multi-line and buffer the next line.

Hope this helps

回答4:

What about storing the dump into a CSV file with mysqldump, using the --tab option like this?

mysqldump --tab=/path/to/serverlocaldir --single-transaction <database> table_a

This produces two files:

table_a.sql that contains only the table create statement; and
table_a.txt that contains tab-separated data.

RESTORING

You can restore your table via LOAD DATA:

LOAD DATA INFILE '/path/to/serverlocaldir/table_a.txt' 
  INTO TABLE table_a FIELDS TERMINATED BY '\t' ...

LOAD DATA is usually 20 times faster than using INSERT statements.

If you have to restore your data into another table (e.g. for review or testing purposes) you can create a "mirror" table:

CREATE TABLE table_for_test LIKE table_a;

Then load the CSV into the new table:

LOAD DATA INFILE '/path/to/serverlocaldir/table_a.txt' 
  INTO TABLE table_for_test FIELDS TERMINATED BY '\t' ...

COMPARE

A CSV file is simplest for diffs or for looking inside, or for non-SQL technical users who can use common tools like Excel, Access or command line (diff, comm, etc...)

回答5:

I'm afraid this won't be possible. In the old MySQL Administrator I wrote the code for dumping db objects which was completely independent of the mysqldump tool and hence offered a number of additional options (like this formatting or progress feedback). In MySQL Workbench it was decided to use the mysqldump tool instead which, besides being a step backwards in some regards and producing version problems, has the advantage to stay always up-to-date with the server.

So the short answer is: formatting is currently not possible with mysqldump.

回答6:

I found this tool very helpful for dealing with extended inserts: http://blog.lavoie.sl/2014/06/split-mysqldump-extended-inserts.html

It parses the mysqldump output and inserts linebreaks after each record, but still using the faster extended inserts. Unlike a sed script, there shouldn't be any risk of breaking lines in the wrong place if the regex happens to match inside a string.

回答7:

Try this:

mysqldump -c -t --add-drop-table=FALSE --skip-extended-insert -uroot -p<Password> databaseName tableName >c:\path\nameDumpFile.sql

回答8:

I liked Ace.Di's solution with sed, until I got this error: sed: Couldn't re-allocate memory

Thus I had to write a small PHP script

mysqldump -u my_db_user -p -h 127.0.0.1 --skip-extended-insert my_database | php mysqlconcatinserts.php > db.sql

The PHP script also generates a new INSERT for each 10.000 rows, again to avoid memory problems.

mysqlconcatinserts.php:

#!/usr/bin/php
<?php
/* assuming a mysqldump using --skip-extended-insert */
$last = '';
$count = 0;
$maxinserts = 10000;
while($l = fgets(STDIN)){
  if ( preg_match('/^(INSERT INTO .* VALUES) (.*);/',$l,$s) )
  {
    if ( $last != $s[1] || $count > $maxinserts )
    {
      if ( $count > $maxinserts ) // Limit the inserts
        echo ";\n";
      echo "$s[1] ";
      $comma = ''; 
      $last = $s[1];
      $count = 0;
    }
    echo "$comma$s[2]";
    $comma = ",\n";
  } elseif ( $last != '' ) {
    $last = '';
    echo ";\n";
  }
  $count++;
}

来源：https://stackoverflow.com/questions/15750535/using-mysqldump-to-format-one-insert-per-line

标签

mysql

mysqldump

mysql-workbench

mysqladministrator