File Formats

A file can arrive on time, with every byte intact, and still be unusable by the application that receives it. The sender may have exported different columns, used another character encoding, or represented dates and amounts differently. Reliable file exchange depends on agreeing on the contents as well as the connection.

With Files.com, you can receive a partner's file, transform its contents into the structure your application needs, and deliver the result through an automated workflow.

This can form part of an extract, transform, load (ETL) process: a source system exports data, a transformation prepares it, and a receiving system loads it. The receiving system confirms whether the delivered records are accepted into the application or database.

For example, a supplier might send a semicolon-delimited order file with decimal commas, while your application expects JSON with numeric amounts. Changing the extension to .json does not perform that conversion. The workflow needs to read the source's representation, map the fields and values, and write the receiving format while preserving identifiers and the relationship between orders and their lines.

Agree on the Data Specification

A data specification describes how a sender represents information and how a receiver interprets it. A sample shows actual records; the specification explains which fields are required and whether a date such as 01/02/2026 means January 2 or February 1.

The agreement needs to cover the choices that affect this exchange:

The sender's export settings and the receiver's import requirements establish the two ends of the transformation. A mapping connects them: which source field supplies each destination field, which conversions are required, and what happens when a required value cannot be interpreted. A default value is appropriate only when the agreement defines what it means.

Identify the Contents

A filename extension tells software which reader to try. A MIME type, also called a media type, labels a kind of content, such as text/csv or application/pdf. Neither label proves that the bytes have that format. A download named orders.csv could contain an HTML sign-in page, for example.

Some formats have recognizable signatures near the beginning of the file. Others, including CSV, depend on the contents following an agreed structure. Identification tools can offer useful evidence, but a ZIP signature alone cannot distinguish an ordinary archive from an XLSX workbook, which is also a package of files.

When a reader reports the wrong file type, inspect the original with a tool appropriate to its contents and check how it was exported or downloaded. Use conversion to write the representation the receiving application needs.

Documents and Records

A document such as a PDF or Word file is arranged for people to read. A structured data file gives software identifiable fields and records. A table that looks orderly on a PDF page may be positioned text rather than a table with machine-readable rows and columns.

The Extract node reads document text and metadata for use in a workflow. That text can support routing or subsequent interpretation. Turning it into a dataset also requires identifying columns, data types, and relationships. An image-only scan first needs optical character recognition (OCR) to turn the image into text; the resulting text still needs interpretation and validation.

TransformScript reads supported structured formats and maps their data into new records. Choose document conversion when the result is another document for people to use; choose a data transformation when the result must satisfy an application's record specification.

Compression and Archives

Compression reduces the representation's size. Archiving packages files, names, and possibly folders into one delivery. They are separate operations: gzip commonly compresses one file, while TAR packages files without compressing them. A .tar.gz file combines both; a ZIP archive packages files and can compress their contents.

The receiving process needs the right operation before it can read the data. Decompressing orders.csv.gz yields a CSV file. Decompressing delivery.tar.gz yields a TAR archive whose members still need to be extracted. Confirm that the receiver can extract TAR as well as decompress gzip when choosing that package.

Files.com's Zip & Unzip guidance explains its archive creation and extraction capabilities. Packaging preserves the enclosed files' formats; it does not convert their records or character encoding.

Parsing and Validation

Parsing turns a file's representation into data a program can work with. Validation checks whether that data satisfies the agreement. A JSON document can parse correctly while omitting a required account identifier; a CSV can have consistent columns while containing amounts in the wrong currency.

A schema is a machine-readable description of structure and constraints. JSON Schema and XML Schema Definition (XSD) are established schema languages. A schema can express rules such as required fields and value types, while business acceptance can also depend on other records or systems.

Files.com Content Validation uses a TransformScript to express the checks required for your delivery. Its verdict comes from those checks. Choosing a file format does not supply the sender's business rules, and a schema document is distinct from the validation script.

For a recurring exchange, test an accepted source file through to acceptance by the receiving application, along with rejected samples that violate important rules. Preserve the received original when you need to investigate a discrepancy or repeat a transformation. The file processing and routing guidance connects these stages to the workflow's operating responsibilities.