Updates to symbology support files for batch downloads

Starting February 17, 2026, we’re changing when symbology support files are included in historical batch downloads and how they're organized.

Currently, these files are provided with every batch download request and map integer instrument IDs to their corresponding raw symbols. In many cases, this information is already present in the data, making the support files redundant and often larger than the data itself. This increases download sizes, slows job completion, and adds unnecessary files.

As a result of this change:

• Symbology support files will no longer be generated for jobs using the DBN encoding because this information is always embedded in the DBN file headers.
• For CSV and JSON downloads, symbology support files will be generated only when symbol mapping is disabled; when symbol mapping is enabled (the default), the symbol field is included, and the support files are redundant.
• When a batch download is split into multiple files, any generated symbology support files will now be created per data file rather than once per job.

This change preserves the existing use cases for symbology support files while reducing unnecessary overhead for most workflows.

DBN batch downloads will no longer generate a symbology support file, since raw symbols are already stored directly in the file headers. You can continue accessing this information using Databento’s client libraries and the DBN CLI tool.

import databento as db

data = db.read_dbn("glbx-mdp3-20251201-20251231.ohlcv-1d.dbn")

print(data.metadata.mappings)

For CSV and JSON, symbol mapping is controlled by the map_symbols parameter and is enabled by default. This means raw symbols are already included directly in each record. In these cases, a separate symbology support file will no longer be generated.

Disabling symbol mapping for CSV or JSON downloads is generally not recommended, but may be useful when you need a separate symbology support file—most notably to maintain compatibility with streaming workflows that use the get_range method. In those cases, you can disable symbol mapping by setting map_symbols=False:

import databento as db

client = db.Historical()

job_details = client.batch.submit_job(
    dataset="GLBX.MDP3",
    symbols=["ESH6"],
    schema="ohlcv-1d",
    start="2025-12-01",
    end="2026-01-01",
    encoding="csv",
    compression="none",
    split_duration="none",
    # disable map_symbols to generate symbology files for CSV and JSON encodings
    map_symbols=False,
)

When a batch download is split into multiple flat files (by size, symbol, or duration), any generated symbology support files will now be created per data file rather than once per batch download. These files will also be named to match the data file they apply to.

Before

glbx-mdp3-20251201-20251231.ohlcv-1d.csv
symbology.csv
symbology.json

After

glbx-mdp3-20251201-20251231.ohlcv-1d.csv
glbx-mdp3-20251201-20251231.symbology.csv
glbx-mdp3-20251201-20251231.symbology.json

You may need to update your workflow before February 17, 2026, if you currently:

• Expect a symbology support file for every batch download request.
• Assume there is exactly one symbology file per download.
• Rely on the generic filenames symbology.csv or symbology.json.