Charts and Analyzer unreal disconnection issue from Database

[pstrusi]

Hi pstrusi,

I'd like to you to clear the chart cache and see if this fixes the chart. (Don't delete the saved historical data though)

Please use the steps I have provided below to delete your chart cache and historical data.

• Close NinjaTrader
• Open (My) Documents\NinjaTrader 7\db\cache\
• Delete all files within this folder
• Restart NinjaTrader
• Connect

Open a new chart without any indicators or templates applied to see if you are able to load a chart.

If this does not fix the chart, please send an email to platformsupport [at] ninjatrader [dot] com so that we may schedule a call and I can assist.

Hi Chelsea,

Following your instructions, I've just performed the "Chart clearing cache and historical data, exactly as you wrote above; unfortunately I could check that the issue remains, the incomprehensible data corrupted error still is shown. So, it appears that we must find another harder reset procedure of this set of data.

I have a doubt remained in my mind. Despite I have my workspace in NT, previously to close the platform, I closed all charts and windows, in order to avoid some stubborn cache data. I wonder, did I have to erase too the " _workplaces " file too?

Hopefully you could come with more suggestions. Looking forward for your post.

[NinjaTrader_ChelseaB]

Hi pstrusi,

I would like to take a look with you on your end.

Please send an email to platformsupport [at] ninjatrader [dot] com. In the email, please include a link to this forum thread.

[pstrusi]

Hi pstrusi,

I would like to take a look with you on your end.

Please send an email to platformsupport [at] ninjatrader [dot] com. In the email, please include a link to this forum thread.

I'll send you the email, but what do you mean exactly by "take a look in my end" ?

[NinjaTrader_ChelseaB]

Hi pstrusi,

I would like to connect with your computer over a remote support session so that I may investigate why the chart does not match the historical data as it appears in the historical data manager.

[pstrusi]

Hi pstrusi,

I would like to connect with your computer over a remote support session so that I may investigate why the chart does not match the historical data as it appears in the historical data manager.

Chelsea, the email is still zipping my chart data so you can check it.

It's necessary that we can find more ways to approach the research, but a remote support session is not possible for me in this pc, due to contractual conditions of my work to firms which I develop strategies, and discretion and security of the Algos is essential. Despite we don't touch codes whatsoever is thought, the remote session implies openning a connection. I hope you understand this. I'm sure that we could continue to workaround logically this evident error.

Please , confirm that you have received the email whenever you got it.

[pstrusi]

I've found out something very important which is the key of the issue.

Going back to the data of 04-05-2016 , I saw something obviously wrong. It turns out that data from 9:00am untill 10:00am, EXTENDS itself beyond 10:00am, right to the moment of the one of the mistaken-events, 10:08:05 ; which obviously explains why NT can't clear correctly the data, cause the datas ource, somehow is messed up in some points. So now we must find out two things:

1. How to clear the bad data ( hopefully it is not a manual procedure )
2. What is the cause that triggers this mess in the database?


I hope this contributes with your search and addressing this sooner. In the email sent by me yesterday, with chart data, you might check for yourself, what I'm telling.

Looking forward for your assistance. Thanks in advance

[NinjaTrader_ChelseaB]

Hello pstrusi,

I was expecting that the historical data itself is the issue. This is why I wanted to connect with you and look at the historical data myself and verify that is correct.

You would need to either manually correct this data, or get it from another source as IB will not have the tick data.

This also confirms that NinjaTrader saved the data as it was received from IB. This may indicate some kind of connection loss or other issue receiving the data from IB.

Historical data is not saved in the database but is instead saved in individual historical data files.

[pstrusi]

Hello pstrusi,

I was expecting that the historical data itself is the issue. This is why I wanted to connect with you and look at the historical data myself and verify that is correct.

You would need to either manually correct this data, or get it from another source as IB will not have the tick data.

This also confirms that NinjaTrader saved the data as it was received from IB. This may indicate some kind of connection loss or other issue receiving the data from IB.

Historical data is not saved in the database but is instead saved in individual historical data files.

Chelsea, I should clear out several important things:

1. After last Friday's close, all the chart data WAS OK because all charts and backtest didn't show any issue. After that I turned my pc off on Saturday night. When on Sunday I start it again my pc and NT to continue with bactest, I've realized that something was wrong. So obviously something happened in the NT database when I restart it again. So the explanation that the data was corrupted by a connection loss or other event during last week is not real. So, WHAT HAPPENED?

2. Regardless there was a connection loss or other issue receiving data ( Which is not the cause of this issue ) , it doesn't justify why NT continues to save data by duplicity. If NT is saving data for the timeframe 09:00am - 10:00am , the new data coming after 10:00 am should be saved only in the new timeframe 10:00am - 11:00am, don't you think so?. Obviously NT should manage to save the data accordingly to the correct time.

Obviously some kind of internal problems are happening in NT7 under Windows 10 at 64bits.

I realize that not only I have to clear my databse manually but also check in that mess if prices are genuine or not. This is something that NT user should not do. Trading markets is a very complicate work and now even more the platform that I trust is becoming a new problem. I hope you understand my frustration here.

[NinjaTrader_ChelseaB]

Hi pstrusi,

I'm not quite understanding your inquiry.

Historical tick data is recorded by each tick and doesn't really have time frames like this. The session template is applied to the tick data after it is loaded.

Each tick will be stamped with the time it is received.

Are you saying there are ticks that are stamped with the wrong time?

Are you saying there are twice as many ticks as expected?

[pstrusi]

Hi pstrusi,

I'm not quite understanding your inquiry.

Historical tick data is recorded by each tick and doesn't really have time frames like this. The session template is applied to the tick data after it is loaded.

Each tick will be stamped with the time it is received.

Are you saying there are ticks that are stamped with the wrong time?

Are you saying there are twice as many ticks as expected?

Exactly that. That's why I've requested to see the data by yourself. There's an inexplicable duplicity in wrong timestamp.

[pstrusi]

Chelsea and whole NT's support and development team.

NT7 working through Windows10 (64bit) presents a CRITICAL FAILURE that is a serious deal-breaker to me. I'm terribly frustrated. This issue or event is happening for quite some time that I've posted previously under other titles. I've had to re-build patiently my data, but then eventually HAPPENS AGAIN !!! .This annoying event has made me lose DATA, TIME and MONEY. The error is simple:

All database, Charts, backtests..etc are OK after Friday's close ( I've proved this always ), then, RANDOMLY EVENTUAL, after close NT and restarted it again on Sunday morning, a big mess is in the database has occurred, basically I've realized that a big bunch of tickdata is stamped OUTSIDE of its natural time range, generating an inexplicable duplicity of data that generates a complete havoc in my platform.

Since I'm not a professional programmer or developer, finding out what's happening exactlyis highly improbable. I'm turning to you cause I need help here. Now the solution is not just repair manually the data or buying new data, because EVENTUALLY in the future this WILL BE AGAIN.

I NEED A PERMANENT SOLUTION TO THIS BEFORE I HAVE TO BUY OR DOWNLOAD NEW DATA

[NinjaTrader_Brandon]

Hello pstrusi,
Thanks for your post.
I am the technical support supervisor here and I have read over your tickets and email with Chelsea.
Please confirm for me which of the following is true.
Are you saying that your historical data looks different when restarting or are you saying that your executions look different after restarting?

[pstrusi]

Hi Brandon,

Thanks for your help.

Yes, as I've told before all was ok until last Friday's close, but after close the whole platform and restarted it again in the next morning, the backtest were terribly different. So after search in the database, I've realized that the Historical looks different therefore any backtest or chart are altered too.

This annoying issue has happened to me in the past a couple of times, but I thought that was some eventual malfunction of my pc or NT, so patiently I uninstalled and installed again from scratch the whole NT and rebuilding the database from IB ( which doesn't have any historical data service to download ). I've done this twice....but enough is enough.

I hope you could find out what's happening

Thanks in advance

[pstrusi]

Brandon. There's an info that might help you in your search about what's causing the issue.

The events have only happened during any week that I've done any live simulation with the account sim101. The issue doesn't rise up right away, just when I close the platform and restart it again. Maybe it's a coincidence, but it happens. Regardless I reset or not the account sim101, it occurs randomly.

[NinjaTrader_Brandon]

Hello pstrusi,
Thanks for your reply.
Please go to Tools> Historical Data Manager> Export
Export Last Tick data for the $EURUSD on 4/5/2016
Attach the file to this thread so I can test this further on my end.