If I update the iqfeed client to 6.0.0.4 and keep all of my existing code running on protocol 5.2 nothing should break, is that correct?

Curious why LabelAtBeginning became the default option for historical interval data? Seems like leaving LabelAtEnd as the default would have been less disruptive and prevented the need for digging through old code to add a field.

I'd like an answer regarding the interval changes as well, why?

Hello,

This change was to bring us in line with what it felt like new developers were expecting from us. Many competitors do not provide an option of how to label their intervals and when comparisons were being done, there would invariably be off by one problems. This was to help eliminate that hurdle.

We felt that, since this would not affect any past protocols, that we could add this change without impacting users much. My expectation was that if someone was going to be editing their code anyhow, to update the protocol, then adding the additional parameter to the HIT, HIX, and HID requests would be acceptable.

My apologies if I underestimated that.

Tim

Perhaps the transition to 6.0 could be made easier if there were a global option that could be set during connection time and would apply a default value to LabelAtBeginning for all subsequent historic data requests (similar to setting the CLIENT NAME). Then users could set that option directly adjacent to the code that sets the protocol, and would not have to worry about modifying any of the data queries elsewhere in their program. Perhaps something such as this:

S,SET OPTION,DefaultLabelAtBeginning,0

Then, if the users use the standard HIX/HID/HIT commands without the extra LabelAtBeginning field (as in legacy 5.0-5.2 code), the new DefaultLabelAtBeginning would apply, and if they do specify the field then it would take precedence.

You can easily extend this mechanism to other options in the future.

---

- Yair
IQML - IQFeed-Matlab connector
https://undocumentedmatlab.com/IQML

Edited by altmany on Aug 2, 2018 at 04:41 AM

That is certainly fair, that would also be useful for setting protocols and more.

We'll discuss it and see if it could be worked in sometime for a future release. Thanks for the idea.

Tim