postgresql

The previous [article] ( /posts/writing-a-on-delete-trigger-in-postgresql/) was about deletion trigger. This article is about update triggers. I had two tables what_did_you_eat and what_was_the_type. On insertion into what_did_you_eat, the what_type column was populated with the food type. Now, if I update the value in what column of a row from what_did_you_eat, I would want the corresponding row in what_was_the_type table to be updated too. For example, if we have the following data:

The previous [article] ( /posts/writing-a-on-insert-trigger-in-postgresql/) was about insertion trigger. I had two tables what_did_you_eat and what_was_the_type. On insertion into what_did_you_eat, the what_type column was populated with the food type. Now, if I delete a row from what_did_you_eat, I would not want a row to remain in what_was_the_type to refer to a non existing row in the what_did_you_eat table. To handle this, I will use the delete trigger. Like before we first write the trigger function.

Here is a simple example of a on insert postgresql trigger. I have a table called what_did_you_eat. It has the following structure - what will contain what was eaten. I have another table called what_was_the_type with the following structure - I want to store the type of food that was eaten in what_type column. And I want to do this to happen on insert into the what_did_you_eat table. So, off we go to write a trigger.

Say we have a table called person with the following structure --------------------------- column| datatype --------------------------- _id | autonumber name | text age | integer --------------------------- We want to insert multiple records into the table using a stored procedure. If we have a stored procedure with say two input variables name and age, then it will take multiple calls. We can do this using a single call if we use a json input and using the json_to_recordset function.

In this article I am going to talk about how to do a classic 3-tier architecture using docker containers. The 3-tiers will be: Frontend tier: This will host the web application. Middle tier: This will host the api, in our case the REST api. Database tier: This will host the database. We want to expose the web application to the outside world. Simultaneously, we want this layer to be able to talk with the middle tier only not the database tier.

I run two instances of postgres servers - one on my local machine and the other one on my do virtual machine. Both of them are on debian. One of them was running 10.x and the other one 9.x version of postgres. I wanted to upgrade them to the latest version - that is version 11. I was pleasantly surprised to see how simple it was. This is what I did:

Say we have a table called person with the following structure --------------------------- column| datatype --------------------------- _id | autonumber name | text age | integer --------------------------- We want to write a stored procedure to return the highest age that we have in the table. We will use OUT param here. CREATE OR REPLACE FUNCTION get_max_age(OUT max_age int) RETURNS int LANGUAGE 'plpgsql' AS $BODY$ BEGIN SELECT MAX(age) INTO max_age FROM person; END $BODY$; That’s it.

Coming from Microsoft SQL Server, I keep on forgetting how to return a resultset from a stored procedure in postgresql. Here is a small sample of how to do it. Needs a bit more code than SQL Server. Note that postgresql does not have stored procedure, they have function. They are equivalent. Say we have a table called person with the following structure --------------------------- column| datatype --------------------------- _id | autonumber name | text age | integer --------------------------- We want to write a stored procedure to return records with age greater than a specified age.