enter image description here
I wasn't able to create a migration even after following doc.
Since the update (I don't know which, exactly), the flag -n was replaced by -o.
So, now you need to run.
typeorm migration:create -o path/to/my/migration
If you are using multiple datasources
migration:create doesn't work with specific datasources, so
I reccomend you to add two scripts in "scripts" session at package.json like the following:
default typeorm (that you will use for non-specific database stuff):
"typeorm": "ts-node-dev ./node_modules/typeorm/cli.js"
your custom typeorm (that you will use for your specific database stuff, like run & revert migrations):
"myTypeorm":"ts-node-dev ./node_modules/typeorm/cli.js -d path/to/my/ormconfig.ts"
Related
I'm trying to figure out how to run an npm script using docker-compose but I only want to run it once (if the data volume hasn't yet been created -- e.g. the VERY first time I docker-compose build && docker-compose up).
The script uses the Sequelize CLI to run a seed file for the database, but if this is run more than once, it'll error in my database because of a duplicate key constraint violation.
This is because I'm using a data volume (so if it's been run before, it's already persisted).
Oh, and this needs to be run after another script has run (the migration script).
So in order:
npm run db:migrate <-- this can run every time docker-compose up is run
npm run db:seed <-- this can only run once as long as the persistent volume hasn't been created
any other scripts can now run (to start my server)
Are there any concepts like this that can be used with docker-compose?
Which database did you use?
In many cases (such as maraidb, mongodb) you can use the directory /docker-entrypoint-initdb.d
Every mounted file wil be executed in alphabetic order if the container starts.
To do your operation only on first start the fist part of your script chould be check if there is already a database or not.
EDIT: Take a look to the doku, which file types are supportet. .sql ans .js schoul work in the most caes, but for npm I'm not shure
I know typeorm migration:run to run all migration files, but some time I want run a file like insert file, or a modified file. But i can't find any doc for that. How to do that thanks.
This is not possible. You can see the docs of the migration with
typeorm migration:run -h
During development you can locally change your connection settings, e.g. change the path to the migration files in your ormconfig.json to match only your file.
"migrations": [
"src/db/migration/**/yourmigration.ts"
]
If you want to seed the database with data you shouldn't use migrations.
Migrations' purpose is to create the database's structure, This is why migrations are executed serially, therefore running a single migration non-serially misses the point.
For database seeding with typeorm you can use typeorm cli query, you can create a script that read the SQL from a file.
For example in a node.js app & bash:
package.json:
"scripts": {
...,
"typeorm": "npx ts-node -r tsconfig-paths/register ./node_modules/typeorm/cli --config ormconfig-migrations.ts",
"db:seed": "bash scripts/db-seed.sh"
}
db-seed.sh:
#!/bin/bash
query=""
for filename in sql/seed/*.seed.sql; do
query=${query}"$(< ${filename})";
done
npm run typeorm query "${query}";
Pay attention there is a bug in this solution, there is always an error thrown from the logger typeorm uses - ignore it. it doesn't relate to the db operation, and does not represent the success of the process.
A migration file that was shared with others shouldn't be modified, in this case, create a new migration.
If that file wasn't shared and is in development, you can run migration:revert as many as you need, then modify and run migration:run again.
For running a specific migration my solution is in ormconfig.js add:
module.exports = {
...
migrations: [process.env.DB_MIGRATION_DIR || 'src/database/migrations/*.ts'],
...
}
and now use DB_MIGRATION_DIR=path/to/migration.ts npm run typeorm migration:run run your specific migration
Another tip for force re-running a migration, but might need some customization for connection string variables - add in your package.json:
{
"scripts": {
...
"migration:rerun": "psql postgresql://$DB_USER#localhost:5432/$DB_NAME -c "delete from migrations where name like '%$SEED_NAME%';" && DB_MIGRATION_DIR='$DB_MIGRATION_DIR' npm run typeorm migration:run"
...
}
...
}
I dont think we can run any migration particular with cli, but we can run it through query runner. please refer the following code:
const connection: Connection = getConnection();
const queryRunner: QueryRunner = connection.createQueryRunner();
//createDatabase16413343434 is class/migration generated from cli
const createDatabase = new createDatabase16413343434();
await createDatabase.up(queryRunner);
One can write custom code to run one migration like inserting a record or just creating/ deleting a single table kind of things. typeorm provides "QueryRunner" which basically run queries in the background. So just create a file with the code given following code to write custom single migration
code to run migration
Is it possible to run only the next migration file with the sequelize-cli?
I have been looking through the docs and help-section of the cli, and it does not appear to be such a feature.
For instance, I have the following when running the sequelize db:migrate:status command;
Loaded configuration file "config/config.js".
Using environment "development".
up 20170301090141-create-something.js
up 20170301133113-create-else.js
up 20170301133821-Update-some-stuff.js
up 20170301135339-create-some-model.js
up 20170307152706-update-some-stuff-two.js
down 20170316142544-create-an-index.js
down 20170421112638-do-some-refactor.js
I would like to only run the 20170316142544-create-an-index.js.
Of course, I can remove all the relevant files. Then I add each migration back one-by-one, running "all" migrations between each one. But this seems so barbaric.
I came here from Google and was not able to find any options in doc. However, the CLI sequelize-cli db:migrate --help shows an option to select a range of migrations you want to run.
They are:
--to Migration name to run migrations until
--from Migration name to start migrations from (excluding)
Apparently, you need to provide the exact filename not the migration name. (Found in source code)
TLDR:
npx sequelize-cli db:migrate --from will-not-be-included.js --to filename.js
I think the question was already answered, but some people still getting confused. I will try to give a straight answer for this specific question:
npx sequelize-cli db:migrate --to 20170316142544-create-an-index.js
In this specific case there's no need to add the option --from, because 20170316142544-create-an-index.js is the next migration to be applied.
But if the question was:
I would like to only run the 20170421112638-do-some-refactor.js
Then you would have to use the option --from to jump all the other migrations until the desired one. In this case:
npx sequelize-cli db:migrate --from 20170316142544-create-an-index.js --to 20170421112638-do-some-refactor.js
The things is, Sequelize only executes those migrations which are not executed yet, Means which is pending.
It means when you run command sequelize:migrate it will only executes those who is still pending.
I think you maked all models and afther that you're trying to run the migrations.
So, I found one solution, i don't know if this is good idea. Sequelize names the migration files with instant actual date. So ,if you change the first string of name for the other wanted file, e.i.
20190717191628-create-team.js
20190717191909-create-player.js
**for**
20190717191909-create-team.js
20190717191628-create-player.js
After that,order all files like you want , and run the db:migrate command.
You can simply change the sequence of files in the migrations folder. As the filename always starts with timestamp so in order to change the sequence of migrations you just need to update the timestamp accordingly by renaming the file.
20170316142544-create-an-index.js. just see which migration ran latest and add rename this migration timestamp with the timestamp just next to the one that executed. add some typos in other ones after it or move them out of the folder for a second and then migrate.
I want to create simple data base model. I'am using postgresql-provider package major:1 minor:1. I've followed instructions to create model. I've added preparations and resource to my Droplet object. Message I receive after running is
No command supplied, defaulting to serve...
Database prepared
Server 'default' starting at 0.0.0.0:8080
Can someone help me with the problem?
With regards to the message No command supplied, defaulting to serve, this is because the binary executable is expecting a 'command'.
vapor run [command]
.build/[configuration]/App [command]
There are a variety of commands available, such as vapor run prepare to run your database preparations, or vapor run serve to begin the HTTP server. You can even add your own commands.
When the executable is run without any commands, it assumes you meant to run the serve command, which is the meaning of your messsage No command supplied, defaulting to serve.
To suppress this, simply use vapor run serve or .build/[configuration]/App serve to run your Vapor project.
Notice how it said in the message Database prepared. That is because all the tables you've specified in your models already exist.
If you've made changes to your models, you'll first need to revert your changes. Vapor has a set of commands just for preparing a database.
vapor run prepare --revert
and
vapor run prepare
The --revert one will run whatever code you've put in the revert function on your models (usually people just delete the table), and then the other command will run the prepare functions and create your models' tables from scratch again.
Adding a runtime dependency to a package through override buildInputs causes the package to rebuild. Is there a simple way to inject runtime dependencies into a package without recompiling?
So basically adding package/bin to PATH and package/lib to LD_LIBRARY_PATH
If I understand correctly that you want to tweak the environment used when a Nix-installed app is run, not the one used when it is built, then a method I know of is as follows below. By using it, you essentially create a wrapper script, which overrides the "default command". So, something similar like creating e.g. a custom ~/bin/vim script, which adds some options/env overrides to the default vim binary, which is called with a "hardcoded original path" inside the script.
One example of it in nixpkgs is how vimutils.vimWithRC overrides vim command with a custom script. For your own use, you could write more or less something like below:
with import <nixpkgs> {};
writeScriptBin "vim" ''
#!/usr/bin/env bash
export PATH=package/bin:$PATH # whatever you like; I've added what you asked for
export LD_LIBRARY_PATH=package/lib:$LD_LIBRARY_PATH
${vim}/bin/vim --my-options "$#"
'';
If you put it in my-vim.nix, you should be able to install it with:
$ nix-env -e vim # REMOVE NORMAL VIM. I think this should be done first to avoid conflict
$ nix-env -i -f my-vim.nix
And hopefully it'll work and "override" the default vim for you.
DISCLAIMER: I haven't actually tested it in this exact form, sorry. Don't have a Nix console handy at this moment, unfortunately.