data_orderbook_start / data_orderbook_end
data_orderbook_start and data_orderbook_end are symbol metadata timestamps that indicate the earliest and latest times for which order book data is available for a symbol. They define the outer bounds of order book coverage, which is critical when you model depth, slippage, or market impact from L2/L3 data.
Why It Matters
Order book data is often the most demanding to collect and normalize, so its coverage may start later and be less uniform than trades or quotes. If you assume order book coverage exists whenever a symbol has overall data coverage, you may end up with missing depth features or biased liquidity comparisons.
How to use these fields
Typical usage:
- Clip order book requests to the intersection of your analysis window and [
data_orderbook_start,data_orderbook_end]. - Build an “order-book-eligible” universe for depth-based research.
- Apply instrument type filtering (
symbol_type) before coverage checks to ensure you are analyzing comparable instruments.
Outer bounds vs depth continuity
A coverage window does not guarantee continuous snapshots or updates. There may be gaps due to outages or illiquid periods. For execution modeling, consider additional filters such as minimum update rates, minimum depth thresholds, and sanity checks on crossed/locked books.
Is order book coverage required if I only use top-of-book?
If you only need best bid/ask, quote data may be sufficient. Order book coverage is required when you need multiple levels, depth, or queue/impact modeling. Choose coverage fields that match your required inputs.
How do order book coverage windows relate to tradable universes?
They are a prerequisite for depth-based tradability metrics. A symbol can be tradable in practice but lack order book coverage in your dataset, which makes depth-based screening impossible. In that case, either switch to quote/trade-based filters or restrict your tradable universe to symbols with order book coverage.
Real-world example
You estimate market impact using order book depth for a set of symbols in 2021. Some venues only started order book collection in late 2021, so data_orderbook_start is after your study start. If you ignore it, you will have missing depth early in the sample and may incorrectly conclude those markets had thin depth.