Hi Guys,
Here's something new for ya!
Have gotten around to using the TimeSales information. Mostly have it figured out, but have run into situations where the historical data gets overwritten with newer information.
I'm asking for 10 days of tick data, and with a clean start of winros and winsig it seems to work fine. But then something happens and the oldest data gets overwritten with todays data, although today's data continues to flow in correctly.
Am sure this is a minor bug to fix. It would seem a simple ASSERT() if the next newer record was older in the history would find it pretty quick.
My algorithm waits until all the history has stabilized coming in before getting anything, and adjusts window correctly to retrieve records 5000 at a clip until all are accumulated.
Here is a snippet showing log of the error. Attached is a full file with lines marked **** showing correct/incorrect behavior.
Thanks! -Jim
Update PXLW DHist 96224 Hist 96433 RT 209
Process First 96432 Last 91432 nHist 96433
-96432: PXLW 08/12/04 10:07:29 AM 7.76 7 Bid | PSE | TrQBid *****
Processed TS PXLW 5000 need update 1
Update PXLW DHist 96224 Hist 96433 RT 209
Process First 91432 Last 86432 nHist 96433
-91432: PXLW 07/30/04 8:45:54 AM 9.54 300 Trade | CSE | TrBid- *****
Processed TS PXLW 10000 need update 1
Update PXLW DHist 96224 Hist 96434 RT 210
Process First 86433 Last 81433 nHist 96434
-86433: PXLW 07/30/04 11:29:25 AM 9.81 15 Ask | NMS | TrQAsk
Processed TS PXLW 15000 need update 1
Update PXLW DHist 96224 Hist 96435 RT 211
Here's something new for ya!
Have gotten around to using the TimeSales information. Mostly have it figured out, but have run into situations where the historical data gets overwritten with newer information.
I'm asking for 10 days of tick data, and with a clean start of winros and winsig it seems to work fine. But then something happens and the oldest data gets overwritten with todays data, although today's data continues to flow in correctly.
Am sure this is a minor bug to fix. It would seem a simple ASSERT() if the next newer record was older in the history would find it pretty quick.
My algorithm waits until all the history has stabilized coming in before getting anything, and adjusts window correctly to retrieve records 5000 at a clip until all are accumulated.
Here is a snippet showing log of the error. Attached is a full file with lines marked **** showing correct/incorrect behavior.
Thanks! -Jim
Update PXLW DHist 96224 Hist 96433 RT 209
Process First 96432 Last 91432 nHist 96433
-96432: PXLW 08/12/04 10:07:29 AM 7.76 7 Bid | PSE | TrQBid *****
Processed TS PXLW 5000 need update 1
Update PXLW DHist 96224 Hist 96433 RT 209
Process First 91432 Last 86432 nHist 96433
-91432: PXLW 07/30/04 8:45:54 AM 9.54 300 Trade | CSE | TrBid- *****
Processed TS PXLW 10000 need update 1
Update PXLW DHist 96224 Hist 96434 RT 210
Process First 86433 Last 81433 nHist 96434
-86433: PXLW 07/30/04 11:29:25 AM 9.81 15 Ask | NMS | TrQAsk
Processed TS PXLW 15000 need update 1
Update PXLW DHist 96224 Hist 96435 RT 211
Comment