I refactored a project and wanted to try out PostgreSQL instead of MySQL. I now want to migrate the table contents of some tables.
The problem is, when I use a select query like this (don't bother about the names, is just an example)
SELECT id AS id_x, name AS name_x, name2 AS name2_x
I want to export the table data and import it into MySQL. The problem is, that the syntax for INSERT INTO is different in MySQL and PostgreSQL. I don't want to export the whole table, because I also changed some parts of the structure, tried to make it more performant etc. So I just want to get the table data, but I need those AS x thing, because the names of the columns have changed
I already found several links on this topic.
I can use mysqldump to dump the table and set the --compatible=name parameter. The problem here is, that I can't add a SELECT statement, right? I can only add a where check.
Then, I could use the mysql command to export the query I want, but mysql doesn't have any compatible parameter. How would I achieve that?
You could consider to create a temporaty table by issuing
SELECT id AS id_x, name AS name_x, name2 AS name2_x FROM oldtable INTO temptable
And then as a second step export the temptable using mysqldump with the --compatible= parameter.
See https://dev.mysql.com/doc/refman/5.7/en/select-into.html
Related
I am not sure if this has been answered before but what I found i am not sure how to make work for me but here is my problem.
I have a database used to keep track of phones for multiple clients. What needs to be done is have a query that can be ran that will run against multiple databases on the same server. each database uses the same table name that I am looking at but the names are slightly different. I came up with this..
INSERT INTO `export db`.exportinfo2
SELECT *
FROM (SELECT * FROM `export db'.tentantnames).users
WHERE name = 'Caller ID:emergency' AND value > 0
What suppose to happen is from a table that has all the database names is is to got to each database and go into the table labeled users and run a where clause on the data then export results to a different database table
I know the code needs to be dynamic but I am not sure how to make it dynamic and function. The table that has all the names for the databases is automatically created every few days.. I am not sure what else needs to be said without sounding like i repeat myself but i just need help making a dynamic query that uses a table premade as database names and run a where statement on the same named table in each database which have their name stored in a different table.
You should look into Synonyms. It can be used to fulfill your purpose
We will migrate the database from mysql to postgresql in our product(through java). So we need to change the mysql query to postgresql query in java application. How to create the table i.e., databasename.tablename in postgresql.
For mysql, we can directly create the table e.g create table information.employee.
Here database name is "information" and table name is "employee" . Is it possible to achieve same query in postgresql.
I searched google it says cross database reference is not possible. Please help me.
I saw pg_class table it contains the table names in the specific database, like wise databse and tables relationships are stored in any other table.
This is normally done using schemas rather than databases, which is more or less like how MySQL organizes it anyway.
Instead of
create database xyz
use
create schema xyz
When you create tables, create them:
create table xyz.myTable
you will need to update your search path to see them on the psql command line tool, or if you want to query them without using the schema explicitly. The default schema is public, so when you create a table without a schema name, it ends up in public. If you modify your search_path as below, the default schema becomes the first in the list: xyz.
set search_path=xyz,public,pg_catalog;
and you must not have spaces in that statement. You can do it globally for a user/role too:
alter role webuser set search_path=xyz,public,pg_catalog;
Also, don't forget that postgresql string matches are case sensitive by default (this one catches people out a lot).
If you want to have different physical locations for the files for each schema, you can do that with tablespaces. If you have a look at the postgresql documentation page, they have info on how to do it, it's pretty easy.
database in MySQL == schema in PostgreSQL. So you will most probably want to migrate all your mysql dbs into one postgres db. Then you will be able to do "cross-database" queries.
See my answer to this question: Relationship between catalog, schema, user, and database instance
I would like to select data from a second MySQL database in order to migrate data from one server to another.
I'm looking for syntax like
SELECT * FROM username:password#serverip.databaseName.tableName
Is this possible? I would be able to do this in Microsoft SQL Server using linked servers, so I'm assuming it's possible in MySQL as well.
You can create a table using FEDERATED storage engine:
CREATE TABLE tableName (id INT NOT NULL, …)
ENGINE=FEDERATED
CONNECTION='mysql://username:password#serverip/databaseName/tableName'
SELECT *
FROM tableName
Basically, it will serve as a view over the remote tableName.
There are generally two approaches you can take, although neither of them sound like what you're after:
Use replication and set up a master/slave relationship between the two databases.
Simply dump the data (using the command line mysqldump tool) from the 1st database and import it into the 2nd.
However, both of these will ultimately migrate all of the data (i.e.: not a subset), although you can specify specific table(s) via mysqldump. Additionally, if you use the mysqldump approach and you're not using InnoDB you'll need to ensure that the source database isn't in use (i.e.: has integrity) when the dump is created.
You can't do this directly, but as someone else alluded to in a comment, you can use mysqldump to export the contents of a table as a SQL script.
At that point you could run the script on the new server to create the table, or if more manipulation of the data is required, import that data into a table with a different name on the new server, then write a query to copy the data from there.
I have a table that has 7000 rows,
I added a new column to this table
The table has a mysql DateTime so.
When i updated the table to fill in this new table it updated the datetime,
I took an sql dump just before i did the update so now i need to use the sql dump to revert the datetime back (and only that column).
How do i do that?
There are a couple ways I can think of to do this off the top of my head.
First is to create another mysql database and load the dump into that database (make sure it's not going to load into the first database from a use commmand in the dump), and then use the data from that database to construct the update queries for the first.
The second, easier, more hackish way, is to open the dump in a text editor, pull out just that table, and find and replace to make update statements for just that column based on primary key instead of inserts. You'd need to be able to find and replace on patterns.
A third way would be to load the dump in an abstract sql tool letting it do the parsing for you, and write new queries from the data in the abstract syntax trees.
A fourth, again hackish, possibility, if this isn't a live system, is to rollback and re-perform the more recent transformations (only if they are simple).
Restore the dump to a second table. Select the ID and datetime from that table. Use those results to update the rows in the original table corresponding to the IDs you got.
Having carefully and painfully manually populated/manipulated a series of records in a table, I want to retain them for reuse. As the table is rewritten daily I'd like to save JUST these particular records as a series of "inserts". The only way I know how to do this is to dump the whole table as sql using a GUI eg sqlyog.
But is there any quicker/better way to do this?
would mysqldump help ? (it's not a GUI)
edit : note that you can save only part of a table using this tool. since it's command line, you can automate the task easily.
Create a copy of your table with a meaningful name and copy using a INSERT the records your interested in.
Doing it this way gives the most flexibility should you need to copy them back/compare them.
Providing you still have your clean manually populated table available. Copy it into another table:
CREATE TABLE mytable_backup SELECT * FROM mytable;
Then you can re-introduce these to your daily rebuild table using a similar method:
INSERT INTO mytable SELECT * FROM mytable_backup;
Does this help?