advice on hyperfile db

Question

At my work, my co-workers are considering using hyperfile as a database server for a windev project. I don't even know that kind of database, it's from PCSOFT, the company that develops windev.

Since windev can also work with microsoft sql server, I'm looking for advice on that kind of database (performance, stability, etc) from people who already used it.

Regards!

Answer 1

Adding upon what Samuël Tremblay already wrote, I would say that after 2+ years of using Windev with HFSQL (old name is HyperFile SQL), here are my conclusions (I have used Windev versions 20 and 22):

PROS:

• replication of a database to another server is rather easy to setup. You can choose to replicate a whole database or a selection of tables. But DBMS like PostgreSQL are actually offering advanced replication setups (https://www.2ndquadrant.com/en/resources/pglogical/).

• easy export to a Microsoft Excel file of a query/table

• create and change the schema/structure of your database through a graphical user interface (GUI)

CONS:

• When you use the database server provided by Windev (i.e. HFSQL), you must use Windev (that is imposed upon you). You cannot interact with your database with another language/framework other than Windev, you are forced to use Windev to query a HFSQL database. If you use instead a DBMS like PostgreSQL, mySQL/MariaDB, etc. you can (and will be able) to query the database with some other language : C++, Java, JavaScript, etc. Say that you wanted now to open your data to customers through a web app, you will actually need to use their other software Webdev from their software suite (and buy it actually). Or say, some day, you want to develop a simple app for smartphone with Qt or else. Well, if your database runs on HFSQL, then you will not be able to query your database unless you use Windev (actually Windev Mobile that you also need to purchase).
• UNIQUE constraints are not working with the presence of NULL (two rows containing NULL would be considered to violate the UNIQUE constraint).
• (almost) every time you update your "analyse/analysis" (basically the database schema), you will also need to update your binary executable. You will need to recompile your software and distribute it again to the users. For example, say you modify a table by adding a column, or modifying the type of a column, then you need to recompile. The executable that the users have will not run, it will say that the version of the "analysis" (schema) on the database is not the same as the one in the executable, and will stop. BAM !
• the HyperFile SQL (HFSQL) server is not so stable, it will crash (often) when executing slightly advanced queries with not so many rows...
• You cannot create scripts to query the HFSQL database : you must create a binary executable (a new project) with Windev. Say you want to quickly modify something --> you need to recompile (and have a Windev IDE with you).
• Say you are on the go, on some trip, and you forgot to bring your computer with the Windev Dongle key (a license cryptographic USB key : it you don't have it, you cannot run Windev), and you need to do some work on the database. PCSoft provides a software called HFSQL Control Center (a GUI software) that can interact with the database, but unfortunately it cannot be downloaded from the internet. You actually obtain it when you buy Windev, and you are allowed to distribute it to whom you want, but it cannot be downloaded from PCSOFT website. Whereas if your database engine is another one, say PostgreSQL or MariaDB, you can simply download PGAdmin or the equivalent, and boom you can interact with your data.
• It seems to me that HFSQL is not a real/genuine DBMS, let me explain myself : the constraints you can set in the analysis (UNIQUE for example), are not always respected. For example, after adding a UNIQUE constraint in the schema (analysis) and compiling the program, I have seen that if I would insert some data in a table from the executable, it would detect the violation of the UNIQUE constraint when it should happen. However if I would insert the same set of data through the HFSQL Control Center, the constraint would not be enforced and the duplicates would be insterted.
• There would be more to say ...

Bottom line: From my own experience, I would strongly encourage anybody, who wants to develop a reliable and dependable software that "must" be developed with Windev (and that needs data persistence), not to use their database HFSQL. You would be much better off using a RDBMS such as PostgreSQL or MariaDB. We are actually going to port our databases from HFSQL to PostgreSQL this Summer.

Answer 2

It depends on the size of your project. Actually, Windev works well with HyperFileSQL. It has been designed for it ! By using another DBMS, you cut yourself some feature such as direct-reading/modifying/deleting in your tables.

Your performances will decrease significantly as soon as you have a nice amount of records in a table (> 100'000). Your database management will become a nightmare since you can't execute several SQL requests at the same time. In example, i'm using another tool developped by a french guy to manage my databases and execute some updates.

Despite of this, it's stable and provides an easy way to interact with Windev's fields.

In my opinion, Hyperfile SQL should be used with small applications with a small amount of features and datas.

Answer 3

You should carefully consider what sql functions you will use. For example deg2rad, rad2deg, ... not working correctly.

Also if you want to use it on a mobile device (Windev Mobile for iOS or Android) you should use SQLLite. Because HyperFile uses a lot of memory and it will be a problem on mobile.

Answer 4

If you want a free database, use PostgreSQL, the Windev connector for PostgreSQL is free to download and install on your windev as a replacement for HFSQL, it will be way more powerful while using the usual hFunctions like you would with HFSQL, plus you will find a ton of docs on the web to do powerful stuff.

HFSQL is in fact the same as an old ISAM DBASE database so it requires re-indexations and things like that of those older DB systems era.

PostgreSQL is like having a free Oracle DB with all the powerful features and reliability, we dropped HFSQL for this and performance has increased tenfold plus all the other benefits while keeping our code pretty much the same, every day feels like we discover freebies and gifts from ProsgreSQL since our migration :)

Free VS Free... You gotta go with power and sheer size of web documentation and poeple available to help .