Login Page - Create Account

Support Board


Date/Time: Fri, 29 Mar 2024 08:35:04 +0000



Post From: Python for Sierra Chart

[2014-07-08 00:32:15]
Kiwi - Posts: 372
Hi Ganz,

Very nice. One thing that has me confused. Why use HDF5?

I did some research after your Python post in the weekend and it seemed to me that HDF5 was about very very big data sets and random access into them. Also the ability to group different data types.

Now Sierra data is essentially sequential in nature and 2 data types (double + int) with fixed data in each column. Also the type of operations I do on them is also sequential - sometimes with some work to coerce the data into that of a higher time period before operation.

In that case would it be better to store them as CSV and/or just .scid files? Possibly the CSV files could be stored with some form of compression. I'm leaping completely out of my experience here and suggesting the bz2 sequential compression ... I probably need to try it.

https://docs.python.org/2/library/bz2.html#sequential-de-compression

Or does HDF5 do really nice compressed serialization?

My other question relates to Python vs C and how the two should fit into the Sierra Chart world so I'll address it in your other thread. I need to think about it a bit more first though.
Date Time Of Last Edit: 2014-07-08 00:40:04