I'm trying to update multiple rows in a DB using a small script.
I need to update the rows based on some specific user_ids which I have in a list on Linux machine.
#! /bin/bash
mysql -u user-ppassword db -e "update device set in_use=0 where user_id in ()";
As you see above, the user_ids are in a file, let's say /opt/test/user_ids_txt.
How can I import them into this command?
This really depends on the format of user_ids_txt. If we assume it just happens to be in the correct syntax for your SQL in statement, the following will work:
#! /bin/bash
mysql -u user-ppassword db -e "update device set in_use=0 where user_id in ($(< /opt/test/user_ids_txt))";
The bash interpreter will substitute in the contents of the file. This can be dangerous for SQL queries, so I would echo out the command on the terminal to make sure it is correct before implementing it. You should be able to preview your SQL query by simply running the following on the command line:
echo "update device set in_use=0 where user_id in ($(< /opt/test/user_ids_txt))"
If your file is not in the SQL in syntax you will need to edit it (or a copy of it) before running your query. I would recommend something like sed for this.
Example
Let's say your file /opt/test/user_ids_txt is just a list of user_ids in the format:
aaa
bbb
ccc
You can use sed to edit this into the correct SQL syntax:
sed 's/^/\'/g; s/$/\'/g; 2,$s/^/,/g' /opt/test/user_ids_txt
The output of this command will be:
'aaa'
,'bbb'
,'ccc'
If you look at this sed command, you will see 3 separate commands separated by semicolons. The individual commands translate to:
1: Add ' to the beginning of every line
2: Add ' to the end of every line
3: Add , to the beginning of every line but the first
Note: If your ID's are strictly numeric, you only need the third command.
This would make your SQL query translate to:
update device set in_use=0 where user_id in ('aaa'
,'bbb'
,'ccc')
Rather than make a temporary file to store this, I would use a bash variable, and simply plug that into the query like this:
#! /bin/bash
in_statement="$(sed 's/^/\'/g; s/$/\'/g; 2,$s/^/,/g' /opt/test/user_ids_txt)"
mysql -u user-ppassword db -e "update device set in_use=0 where user_id in (${in_statement})";
im trying to detect and delete a line break out of a subject (called m.subject) mail information retrieved via CONCAT out of a mysql database.
That said, the linebreak may or may not occur in the subject and therefore must be detected.
My query looks like this:
mysql --default-character-set=utf8 -h $DB_HOST -D $TARGET -u $DB_USER -p$DB_PW -N -r -e "SELECT CONCAT(m.one,';',m.two,';',m.three,';',m.subject,';',m.four';',m.five,';',(SELECT CONCAT(special_one) FROM special_$SQL_TABLE WHERE msg_id = m.six ORDER BY time DESC LIMIT 1)) FROM mails_$SQL_TABLE m WHERE m.rtime BETWEEN $START AND $END AND m.seven = 1 AND m.eight IN (2);"
I tried to delete it afterwards, but getting in performance trouble due to several while operations on all lines already. Is there an easy way to detect and cut it directly via the CONCAT buildup? It is crucial to retrieve only one line after extraction for me.
Updating/changing the database is not an option for me, as I only want to read the current state.
I have a query that takes a very long time to complete. I am wondering if there is anyway to optimize it so it would run quicker. Currently the table has around 15 million row and has indexing on expirationDate. This query has been running for 8000+ seconds.
UPDATE LOW_PRIORITY items
SET expire = 1
WHERE expirationdate < Curdate()
AND expirationdate != '000-00-00'
AND expire = '0'
AND ( `submit_id` != '742457'
OR submit_id IS NULL )
Never mind your table structure (I requested it before this edit), I have a solution for you. I've done this before with my DBAs several times on our production mySQL databases even during peak hours of the day. The approach you need to take is taking your one massive UPDATE query (that will lock the table while executing) and turn it into several individual updates. Those individual updates will not have a material impact on your database, especially if you run them in smaller batches and break it down.
To generate those individual queries, run this (assuming the items table's primary key is named "id"):
SELECT CONCAT(
‘UPDATE items SET expire = 1 WHERE id = ‘,
items.id,
‘;’) AS ''
FROM items
WHERE expirationdate < Curdate()
AND expirationdate != '000-00-00'
AND expire = '0'
AND ( `submit_id` != ‘742457'
OR submit_id IS NULL )
This will generate SQL that looks like this:
UPDATE items SET expire = 1 WHERE id = 1234;
UPDATE items SET expire = 1 WHERE id = 2345;
.....
What I personally like to do is put the above query in a text file and run this from the command line:
cat theQuery.sql | mysql -u yourusername [+ command line args with db host, etc] -p > outputSQLCommands.sql
This will pass the query to mySQL and dump the results to an output file. If you notice, we named the column in our output to '' (nothing). That makes the output very nicely formatted in the text file so it can be easily dumped back into mySQL to run the commands. In fact you could do it all in one fell swoop like this:
cat theQuery.sql | mysql -u yourusername [+ command line args with db host, etc] -p | mysql -u yourusername [+ command line args with db host, etc] -p
Then you are definitely ballin' SQL style.
After you have the output SQL file though, if it's too many queries to run one after another, you can use a command line tool like "split" to split the file into a bunch of smaller files.
I would recommend testing to find the optimum batch size based on your DB's hardware/performance/etc. For example, try running 1000 updates first, if that works with no material impact try 5000, etc. Then you can split the rest of your file into chunks you are comfortable with and run them through.
Good luck!
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?
Try use the following option:
--skip-extended-insert
It worked for me.
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.
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
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...)
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.
Try this:
mysqldump -c -t --add-drop-table=FALSE --skip-extended-insert -uroot -p<Password> databaseName tableName >c:\path\nameDumpFile.sql
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.
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";
}
$count++;
}
add
set autocommit=0;
to first line of your sql script file, then import by:
mysql -u<user> -p<password> --default-character-set=utf8 db_name < <path>\xxx.sql
, it will fast 10x.
I've been trying to get a shell(bash) script to insert a row into a REMOTE database, but I've been having some trouble :(
The script is meant to upload a file to a server, get a URL, HASH, and a file size, connect to a remote mysql database, and insert the data into an existing table. I've gotten it working until the remote MYSQL database bit.
It looks like this:
#!/bin/bash
zxw=randomtext
description=randomtext2
for file in "$#"
do
echo -n *****
ident= *****
data= ****
size=` ****
hash=`****
mysql --host=randomhost --user=randomuser --password=randompass randomdb
insert into table (field1,field2,field3) values('http://www.example.com/$hash','$file','$size');
echo "done"
done
I'm a total noob at programming so yeah :P
Anyway, I added the \ to escape the brackets as I was getting errors. As it is right now, the script is works fine until connects to the mysql database. It just connects to the mysql database and doesn't do the insert command (and I don't even know if the insert command would work in bash).
PS: I've tried both the mysql commands from the command line one by one, and they worked, though I defined the hash/file/size and didn't have the escaping "".
Anyway, what do you guys think? Is what I'm trying to do even possible? If so how?
Any help would be appreciated :)
The insert statement has to be sent to mysql, not another line in the shell script, so you need to make it a "here document".
mysql --host=randomhost --user=randomuser --password=randompass randomdb << EOF
insert into table (field1,field2,field3) values('http://www.site.com/$hash','$file','$size');
EOF
The << EOF means take everything before the next line that contains nothing but EOF (no whitespace at the beginning) as standard input to the program.
This might not be exactly what you are looking for but it is an option.
If you want to bypass the annoyance of actually including your query in the sh script, you can save the query as .sql file (useful sometimes when the query is REALLY big and complicated). This can be done with simple file IO in whatever language you are using.
Then you can simply include in your sh scrip something like:
mysql -u youruser -p yourpass -h remoteHost < query.sql &
This is called batch mode execution. Optionally, you can include the ampersand at the end to ensure that that line of the sh script does not block.
Also if you are concerned about the same data getting entered multiple times and your rdbms getting inconsistent, you should explore MySql transactions (commit, rollback, etc).
Don't use raw SQL from bash; bash has no sane facility for sanitizing the data beforehand. Generate a CSV file and upload that instead.