Skip to main content

PC-Axis (PX) Format

Description​

PC-Axis (PX) is a statistical data format developed by Statistics Sweden and used by many statistical offices worldwide. It's a text-based format designed for exchanging multi-dimensional statistical data with metadata. The format supports hierarchical data structures with dimensions (stub and heading variables) that are flattened into tabular records.

File Extensions​

  • .px - PC-Axis statistical data files

Implementation Details​

Reading​

The PC-Axis implementation:

  • Parses metadata section (TITLE, STUB, HEADING, VALUES, CODES, etc.)
  • Parses data section (numeric values)
  • Flattens multi-dimensional data into tabular records
  • Converts numeric values to appropriate types (int/float)
  • Preserves metadata as record fields (prefixed with _)
  • Supports various character encodings

Writing​

Writing is not currently supported for PC-Axis format.

Key Features​

  • Multi-dimensional data: Handles statistical data with multiple dimensions
  • Metadata preservation: Includes metadata fields in output records
  • Automatic flattening: Converts multi-dimensional arrays to flat records
  • Totals support: Can count total records
  • Encoding support: Handles various character encodings
  • Statistical format: Designed for official statistics

Usage​

from iterable import open_iterable

# Basic reading
source = open_iterable('data.px')
for row in source:
    print(row)
    # Each row contains dimension variables and a VALUE field
source.close()

# With specific encoding
source = open_iterable('data.px', iterableargs={
    'encoding': 'iso-8859-1'  # Common for European statistical data
})

# Check total records
source = open_iterable('data.px')
total = source.totals()
print(f"Total records: {total}")
source.close()

Record Structure​

Each record contains:

  • Dimension variables: Fields from STUB and HEADING declarations (e.g., Age, Gender, Year)
  • VALUE: The numeric data value for that combination of dimensions
  • Metadata fields: Additional metadata prefixed with _ (e.g., _TITLE, _CONTENTS)

Example record:

{
    'Age': '0-4',
    'Gender': 'Male',
    'Year': '2020',
    'VALUE': 100,
    '_TITLE': 'Population Statistics',
    '_CONTENTS': 'Number of persons'
}

Parameters​

  • encoding (str): Character encoding (default: utf-8)

Limitations​

  1. Read-only: PC-Axis format does not support writing
  2. Flat data only: Only supports tabular/flat data structures
  3. Multi-dimensional flattening: Complex multi-dimensional data is flattened into records
  4. File format: Requires proper PC-Axis file structure
  5. Encoding: May require specific encoding for non-ASCII data

Compression Support​

PC-Axis files can be compressed with all supported codecs:

  • GZip (.px.gz)
  • BZip2 (.px.bz2)
  • LZMA (.px.xz)
  • LZ4 (.px.lz4)
  • ZIP (.px.zip)
  • Brotli (.px.br)
  • ZStandard (.px.zst)
  • Snappy (.px.snappy)
  • LZO (.px.lzo)

Use Cases​

  • Official statistics: Reading data from statistical offices
  • Government data: Many government agencies use PC-Axis format
  • Data analysis: Converting statistical data for analysis
  • Data migration: Converting from PC-Axis to other formats
  • Research: Working with official statistical datasets
  • CSV - Simple text format
  • SAS - Statistical analysis format
  • Stata - Statistical software format
  • SPSS - Statistical package format