Bar data N days back

sspy

I want to access the data N days ago. I have written the following code

    def next(self):
        print('next:: current period:', len(self))
        print('date %s, current price %.2f, previous price %.2f' % (self.datas[0].datetime.datetime(), self.data.close[0], self.data.close[-1]))

the dataset I use consists of 3 lines

2018-01-01,50.1,...
2018-01-02,50.2,...
2018-01-03,50.3,...

and the output is

next:: current period: 1
date 2018-01-01 23:59:59.999989, current price 50.10, previous price 50.30
next:: current period: 2
date 2018-01-02 23:59:59.999989, current price 50.20, previous price 50.10
next:: current period: 3
date 2018-01-03 23:59:59.999989, current price 50.30, previous price 50.20

How can day 1 previous price be 50.30? Why it takes the value of the last day and doesn't throws an error?

backtrader

Because the data is preloaded and the magic of array arithmetic in Python does the rest ...

@sspy said in Bar data N days back:

I want to access the data N days ago. I have written the following code

You may want to do it better: Docs - Platform Concepts - See: Lines - Delayed Indexing

sspy

@backtrader said in Bar data N days back:

Because the data is preloaded

Can this behavior be disabled? It really confuses me.

backtrader

@sspy said in Bar data N days back:

Can this behavior be disabled? It really confuses me.

You are simply doing things wrong. But you can disable preloading: Docs - Cerebro

sspy

@backtrader said in Bar data N days back:

You are simply doing things wrong. But you can disable preloading: Docs - Cerebro

Thank you for your prompt reply. Finally I was able to get it right.

The only moment by disabling preloading I was able to get

next:: current period: 1
date 2018-01-01 23:59:59.999989, current price 50.10, previous price 50.10

it's certainly better, but still not how it should

sspy

I'm completely confused. Is there a simple way to get data N days back?

backtrader

It would seem you are unaware of the Python arithmetic with arrays.

As quoted above ..

@backtrader said in Bar data N days back:

You may want to do it better: Docs - Platform Concepts - See: Lines - Delayed Indexing

Using a delayed index provides the proper automatic constraint to the platform.

You really want to read the document linked above and also Docs - Operating the Platform

sspy

@backtrader said in Bar data N days back:

It would seem you are unaware of the Python arithmetic with arrays.

I understand the python arrays but this logic is unacceptable here, because we can't get future data, so the -1 has to be the previous day, -2 the day before it and so on and if no such day is available to throw an exception, but not start everything in a circle.

@backtrader said in Bar data N days back:

Using a delayed index provides the proper automatic constraint to the platform.

Can you show me an example of proper way to get data for 2 days back or exeption if no such data is available?

sspy

I was able to find a solution,

self.data.close.get(size=1, ago=-1)[0]

works right as I want