When I update my EOD Data from EODHD I get 3/4 of the time the error:
One or more errors occurred. (Response status code does not indicate success: 429 (Too Many Requests).)
It updates only 1/4 of the symbols. I have to start the update up to 10 times to update all my symbols. Seems to me the the rate limiting in WL is not working with me. Others seem not to have that problem as I found only post some Years ago with similar problems...
As I am a WL Beginner: Is there any Feature or button to avoid these error messages and get a clean update process?
One or more errors occurred. (Response status code does not indicate success: 429 (Too Many Requests).)
It updates only 1/4 of the symbols. I have to start the update up to 10 times to update all my symbols. Seems to me the the rate limiting in WL is not working with me. Others seem not to have that problem as I found only post some Years ago with similar problems...
As I am a WL Beginner: Is there any Feature or button to avoid these error messages and get a clean update process?
Rename
Context needed.
A free API Token is limited to 20 requests per day, but paid plans are allowed 1000 requests per minute. Which do you have?
A free API Token is limited to 20 requests per day, but paid plans are allowed 1000 requests per minute. Which do you have?
I have the max plan (all in once) with 100.000 token. I have no such problems, when I use my self written programs or the programs provided by EODHD. Today, for example, I have still 85.000 tokens available. But I have this problem every day in WL with the EODHD data,
I just tried it on my WL trading machine, where no previous EODHD data were downloaded at that computer. The same errors, but it seems to load half the data without error messages, not only 1/4 of it.
It seems that my API calls are all counted, even the ones that get an 429 error as well (at least its my impression, I have not verified by counting it...)
I just tried it on my WL trading machine, where no previous EODHD data were downloaded at that computer. The same errors, but it seems to load half the data without error messages, not only 1/4 of it.
It seems that my API calls are all counted, even the ones that get an 429 error as well (at least its my impression, I have not verified by counting it...)
It doesn't look like like WL's EOD provider uses throttling. That can be added pretty easily.
What are you updating exactly? I assume a DataSet with many more than 1000 symbols? How many?
What are you updating exactly? I assume a DataSet with many more than 1000 symbols? How many?
From my memory, EODHD plans still have daily limits. Check your dashboard when you log into EODHD. You're hitting a daily limit. Click API Usage.
thank you for your promt answers on the weekend...
@Cone
Yes, I update about 6300 US Stocks every day. The first 200 or so run smoothly, then every 2 or 3 stocks bring this error message. The strange thing is: When I use my own software to download this stock, I don't use throttling and it works just fine with 16 requests in parallel (haven't tried more requests at once yet). But the official EODHD download app is not as fast in downloading as my personal app, which is in python and not optimized. That might be a hint, that they use throttling in their own downloading app.
@glitch
Its definitly not an api token problem. I download these stocks since more than a year every day with the official EODHD app for my live trading with another trading software without any problems. It costs about 13.000 api tokens per run. If it would be a token problem, it would not help to start the data update prozess in WL again. But in WL it helps. But I need to do that about 5 - 6 times to catch all stocks, which then cost 13000 tokens per run. For my tests today I spend 85.000 tokens (thats the reason why I did it on the weekend). So it is very "costly" in term of tokens and its not possible to do automated. And very error prone...
I'm just astonished that nobody else runs into that problem. I'm sure, I'm not the only one, who trades EOD (but perhaps with EODHD). All the other postings I found regarding EODHD were intraday...
@Cone
Yes, I update about 6300 US Stocks every day. The first 200 or so run smoothly, then every 2 or 3 stocks bring this error message. The strange thing is: When I use my own software to download this stock, I don't use throttling and it works just fine with 16 requests in parallel (haven't tried more requests at once yet). But the official EODHD download app is not as fast in downloading as my personal app, which is in python and not optimized. That might be a hint, that they use throttling in their own downloading app.
@glitch
Its definitly not an api token problem. I download these stocks since more than a year every day with the official EODHD app for my live trading with another trading software without any problems. It costs about 13.000 api tokens per run. If it would be a token problem, it would not help to start the data update prozess in WL again. But in WL it helps. But I need to do that about 5 - 6 times to catch all stocks, which then cost 13000 tokens per run. For my tests today I spend 85.000 tokens (thats the reason why I did it on the weekend). So it is very "costly" in term of tokens and its not possible to do automated. And very error prone...
I'm just astonished that nobody else runs into that problem. I'm sure, I'm not the only one, who trades EOD (but perhaps with EODHD). All the other postings I found regarding EODHD were intraday...
@Glitch - a throttle is required to avoid the issue. EODHD advises to "spread out the requests more or less evenly.. without making them go off all at once, to avoid a "Too Many Requests" error.
.. already testing it.
.. already testing it.
Build 17 is available now with the throttling. Let us know if that works for you!
Thank you for the fast implementation. It works for me now.
Now that I don't get the error 429 any more, I have another problem (but thank you forf solving the 429 Problem so quick).
Now every time I try to make a backtest with the EODHD Data, it tries to fetch new data before the backtest. I like to test (and trade) the complete US stocks Univers inclusive the delisted stocks. These are about 24.000 stocks (of whom only 6200 are active). So this look up takes about 30 to 45 min for the EOD data. In 99% of the cases I don't need the new data (and 2/3 of the stocks are inactive, but I need to test them for the bias free testing), If I only want to backtest, it doen't matter if there is a day more or less data. But even more important, this drains my 100.000 API calls a day. So every backtest, where I change the date costs 24.000 API calls and a lot of time for the fetching of data for inactive stocks. If there are intraday data involved, every backtest costs more than 100.000 API calls. Not counted the API costs for Splits, Dividends or fundamental data, which are tenfold. The only solution I found yet, is to delete for the backtests the API key. BUT even then WL tries to fetch the data and it takes the same amout of time to start the backtest (and the thousands of error messages), but at least it doesn't drain my API Call unneccesary.
Is there a solution for that problem, which I haven't found yet? Otherwise it would be nice to have a button, to not search for new data, only use the data which is already downloaded and cached. This would also be extremly helpful as well for the other data sources, where the same problem occurs, only that they don't have a limited number of API calls a day.
Now every time I try to make a backtest with the EODHD Data, it tries to fetch new data before the backtest. I like to test (and trade) the complete US stocks Univers inclusive the delisted stocks. These are about 24.000 stocks (of whom only 6200 are active). So this look up takes about 30 to 45 min for the EOD data. In 99% of the cases I don't need the new data (and 2/3 of the stocks are inactive, but I need to test them for the bias free testing), If I only want to backtest, it doen't matter if there is a day more or less data. But even more important, this drains my 100.000 API calls a day. So every backtest, where I change the date costs 24.000 API calls and a lot of time for the fetching of data for inactive stocks. If there are intraday data involved, every backtest costs more than 100.000 API calls. Not counted the API costs for Splits, Dividends or fundamental data, which are tenfold. The only solution I found yet, is to delete for the backtests the API key. BUT even then WL tries to fetch the data and it takes the same amout of time to start the backtest (and the thousands of error messages), but at least it doesn't drain my API Call unneccesary.
Is there a solution for that problem, which I haven't found yet? Otherwise it would be nice to have a button, to not search for new data, only use the data which is already downloaded and cached. This would also be extremly helpful as well for the other data sources, where the same problem occurs, only that they don't have a limited number of API calls a day.
That can't be the case with Daily data. When a symbol has updated in cache, it's not requested again until it determines that the cached data is no up-to-date. It's simply reloaded from cache.
But you said, "If there are intraday data involved..."
If you're testing during market hours (or anytime with the market filter disabled), then sure, it's going to try to get updated data with every backtest.
Just try File > Offline Mode when you're testing.
But you said, "If there are intraday data involved..."
If you're testing during market hours (or anytime with the market filter disabled), then sure, it's going to try to get updated data with every backtest.
Just try File > Offline Mode when you're testing.
Your Response
Post
Edit Post
Login is required