Local Import

What this tab is for

Use Local Import to turn supported files or folders on this workstation into a Dataset. Arizmic reads the source and writes managed data of its own; it does not rename, sort, or rewrite the original files.

The import flow

1. Choose the source type, format, local path, and Dataset name.
2. Inspect the source. Arizmic reads a bounded sample and proposes the data shape, field mappings, market identity, coverage, and storage layout.
3. Review each proposed mapping and fill only the missing market facts.
4. Validate a sample. Resolve every blocking issue before queueing the full import.
5. Queue the job, then review its candidate Dataset in Catalog before making it active.

Sample validation is deliberately fast. The full job still reads the complete source, so it can find a problem that was not present in the sample.

Source details

• Source type — choose file for one source file or folder when the import belongs across a collection of files.
• Source format — tells Arizmic how to read the source. Select the format that produced the file, not the format you expect the data to become.
• Dataset name — a local label used in Catalog. It does not replace the vendor symbol, contract identity, or source lineage.
• Path or Folder Path — the original local source. Use the picker when possible so the path is exact.

For a folder, inspection also reports how many files are importable, the files it skipped, the detected root symbol, venue, contracts, and any identity conflicts. Do not treat an inspected date range as proof that every instant in that range is present; validation checks the mapped time fields more deeply.

Market identity

The fields under Market Identity describe what the prices mean. They are not cosmetic labels.

• Asset — the market family, such as equity, future, option, forex, or crypto. It determines which instrument facts apply.
• Venue — the exchange or market venue associated with the records.
• Instrument — the specific market being imported. Use the dated contract for futures when the file is contract-specific; do not substitute a root symbol without knowing how the source was constructed.
• Tick Size — the smallest permitted price movement.
• Contract Size — the number of underlying units represented by one contract or lot.
• Tick Value — the cash value of one tick. This and Contract Size affect costs, sizing, and PnL calculations.
• Timezone — the timezone used by source timestamps when they are not already unambiguous UTC values.
• Calendar — the market schedule used to derive sessions and market windows. It is not merely the timezone.

Use the source's data dictionary, export settings, or companion definition file first. For exchange-traded products, confirm missing tick economics, contract size, expiry, and session rules in the exchange's official contract specification. Provider documentation, a broker's contract-details screen, and an official venue calendar are useful cross-checks. Do not estimate these facts from a chart or from a similar symbol. If you cannot establish the source timezone or market identity, stop before validation rather than guessing.

Map the records

After inspection, Detected Output identifies the record shape. Select a different shape only when it describes the actual rows. A fixed-format source may have its mapping locked; in that case, confirm the Market Identity instead of trying to remap a vendor-defined format.

The mapping table connects a source column to the canonical field that it actually represents. Map a value by its meaning, not by a familiar column name.

Time fields

• Event Time is when the market event occurred.
• Received Time is when the feed or vendor received it, when available.
• Start Time and End Time describe the bounds of a bar.

Event and received time are not interchangeable. For a bar, use its published open and close boundaries; do not use the file creation time. If a timestamp is a number, confirm its unit in the vendor's export documentation. If it has no offset or timezone, use the source documentation or export setting to identify the local market timezone before mapping it.

Prices and quantities

• Open, High, Low, Close are the first, highest, lowest, and last traded prices for a bar.
• Price is the price of an individual trade or book update.
• Size is the executed trade quantity or the displayed quantity for the record type.
• Bid and Ask fields are the best available buy and sell prices and their displayed sizes at the quote time.

Do not map volume to trade count, a price to a tick value, or a vendor's notional field to quantity without the vendor defining it that way.

Required fields by record shape

• Bars need their time bounds and Open, High, Low, and Close. Volume and trade count are used when the source provides them.
• Trades need Event Time, Price, and Size. A trade ID, source row, aggressor side, or exchange side improves traceability when available.
• Quotes need Event Time plus Bid Price, Bid Size, Ask Price, and Ask Size.
• Depth Deltas need Event Time, book action, side, price, and size. Map a level or sequence only when the source defines it.
• Order Book Events use the provider's event semantics. Preserve action, side, price, size, sequence, order identifier, snapshot marker, and level whenever the source supplies them; do not invent a missing order-book state.

Use the vendor's schema or data dictionary to resolve ambiguous fields. A companion instrument-definition file can supply symbols and contract details, but it does not prove the meaning of a price or timestamp column in a separate data file.

Validate, queue, and review

Inspection reports the source sample, columns, detected shape, proposed coverage, mapping issues, Instrument resolution, and storage layout. Validate Sample checks the mapped fields and reports the rows checked, normalized fields, and any blockers. Fix the first blocker, inspect again if the source or identity changed, then validate again.

Queue the import only when validation says it is ready. A successful job publishes a candidate Dataset. In Catalog, confirm coverage, issues, and the effect of making it active. Choose New Version for a material replacement; use Append Coverage only when the new records are compatible with the existing Dataset.