Misaligned data

dee

Hi,

I am using the historical data API to query GAIL(instrument id - 136231684) at 60minute interval.

As highlighted below, some of the data I get is misaligned (refer line 1 and line 11 below). The timestamp is of 10:12 rather than 10:15 which should have been the expected 60 minute interval.

This is a serious issue as when I try to fetch data for multiple instruments, each instrument has its own misaligned pieces and as a result, combining them into a single panel becomes impossible.

Could you please investigate.


x=kite.historical('136231684','2017-09-01','2017-10-26','60minute')
>>> pd.DataFrame(x)
close date high low open volume
0 384.55 2017-09-01T09:15:00+0530 386.95 378.55 378.75 40229
1 388.00 2017-09-01T10:12:00+0530 389.55 384.15 384.55 30377
2 387.50 2017-09-01T11:15:00+0530 388.10 385.20 388.00 7036
3 385.40 2017-09-01T12:15:00+0530 388.75 384.25 387.50 9883
4 385.65 2017-09-01T13:15:00+0530 386.10 384.85 385.40 1531
5 384.50 2017-09-01T14:15:00+0530 386.10 384.05 385.65 11003
6 383.90 2017-09-01T15:15:00+0530 384.80 383.00 384.00 8113
7 385.50 2017-09-04T09:15:00+0530 386.45 382.00 384.10 42253
8 385.50 2017-09-04T10:15:00+0530 385.55 383.00 385.50 3601
9 380.55 2017-09-04T11:15:00+0530 385.50 379.00 385.50 10066
10 379.75 2017-09-04T12:15:00+0530 380.55 379.50 380.55 15662
11 379.75 2017-09-04T13:09:00+0530 380.35 378.80 379.75 2155
12 379.80 2017-09-04T14:15:00+0530 380.75 379.25 379.75 14185

Regards,
-Deepak

yousuf

There are lots of errors of this sort and i have also reported them. These errors are making the historical API completely useless. I too am waiting a response....

Vivek

@dee @yousuf We have forwarded this query to our data team and they are investigating this.

yousuf

hi @Vivek,

Even the most carefully written software has bugs, but these kind of errors, and their large number in basic data, simply shows the lack of basic software testing and no quality control whatsoever, with something as critical as data, which acts as input to all of your clients decision making on their trades..! I am really surprised by the number of errors in data, and the amount of time these errors have gone unnoticed in spite of the data being used everywhere, from charting to trading and reporting. My confidence in zerodha is definitely shaken.

dee

@Vivek could you please provide some update. As a paying customer, considering you do charge us separtely for historical data APIs, we do expect some quality. As @yousuf mentioned, these are quite basic errors which IMHO, should be easy to catch at the testing level. Considering the importance of the data stream in all sorts of application(charting, backtesting, optimizing algortihms) , the customers would certainly expect a better turn around time.

sujith

@dee,
It should be resolved by now. Can you check and let us know if you are still facing the issue?