dbfread

You can control whether records are streamed from disk or held in memory using the load() and unload() methods and the load argument during initialization.

• Streaming (Default): When load=False, the records and deleted attributes are RecordIterator objects that stream data from the disk. This is memory-efficient for large files.
• In-Memory: When load=True, or after calling .load(), the records and deleted attributes become lists. This allows for faster access but consumes more RAM.

Use .unload() to clear records from memory and return to a streaming RecordIterator state.

dbf = DBF('data.dbf', load=False)

# Load everything into memory as lists
dbf.load()
print(len(dbf.records))

# Unload to save memory and return to streaming mode
dbf.unload()

In version 2.0, the DBF class is no longer a subclass of list. While iteration and record counting (e.g., len(table)) still work as expected, direct list operations like indexing (e.g., table[1]) will no longer work on a DBF instance.

To access specific records by index, use the .records attribute.

# Old way (broken in 2.0+ if using DBF class)
table = dbfread.read('people.dbf')
print(table[1])

# New way
from dbfread import DBF
table = DBF('people.dbf', load=True)
print(table.records[1])

The dataset package can be used to easily move data from a DBF file into a modern database like SQLite. This method also handles schema creation automatically.

import dataset
from dbfread import DBF

db = dataset.connect('sqlite:///example.sqlite')
table = db['people']

for record in DBF('example.dbf'):
    table.insert(record)

If a DBF field type character is not a valid Python identifier (e.g., + for autoincrement), you must name your parsing method using its hexadecimal ASCII value prefixed with parse.

For example, to implement a parser for +, name the method parse3F (since ord('+') is 0x3F).

You can generate the method name programmatically using:

'parse' + format(ord('?'), 'x').upper()

Replace '?' with your target field type character.

>>> 'parse' + format(ord('?'), 'x').upper()
'parse3F'

To convert a DBF table into a Pandas DataFrame, you must wrap the DBF object in iter(). Without iter(), Pandas will not recognize that the table is iterable.

Note: Pandas creates an internal list of all records before converting them to a DataFrame, which means the entire dataset will be loaded into memory.

import pandas as pd
from dbfread import DBF

df = pd.DataFrame(iter(DBF('example.dbf')))
print(df)

As of version 2.0, dbfread.open() and dbfread.read() are deprecated and scheduled for removal in version 2.2. To ensure future compatibility, you should switch to using the DBF class directly.

Note that open() and read() currently return a DeprecatedDBF object (which is a subclass of both DBF and list for backward compatibility), but you should transition to DBF to avoid breaking changes when they are removed.

# Deprecated approach
table = dbfread.read('people.dbf')

# Recommended approach
from dbfread import DBF
table = DBF('people.dbf', load=True)

Use the DBF class to read DBF files (used by dBase, Visual FoxPro, and FoxBase+). The library iterates through the file and returns each record as a Python dictionary containing native Python data types (e.g., datetime.date for date fields). This is ideal for batch processing and one-off scripts.

from dbfread import DBF

for record in DBF('people.dbf'):
    print(record)
# Output example:
# {'NAME': 'Alice', 'BIRTHDATE': datetime.date(1987, 3, 1)}

You can export DBF data to a CSV format using Python's standard csv module. This involves iterating over the DBF table object and writing each record to a CSV file.

import csv
from dbfread import DBF

for record in DBF('example.dbf'):
    print(record)

# To write to a CSV file using the standard csv module:
with open('output.csv', 'w', newline='') as f:
    writer = csv.DictWriter(f, fieldnames=['NAME', 'BIRTHDATE'])
    writer.writeheader()
    for record in DBF('example.dbf'):
        writer.writerow(record)

To use dbfread, you must have one of the following Python versions installed:

• Python 3.6 or newer
• Python 2.7

To read a DBF file, instantiate the DBF class with the file path. You can iterate over the resulting object to get records as dictionaries (or collections.OrderedDict in older Python versions). Records are returned in the order they appear in the file.

To access deleted records, use the .deleted attribute. You can also use the with statement for convenience, though the DBF object does not keep files open by default.

from dbfread import DBF

# Open the table
table = DBF('people.dbf')

# Iterate over records
for record in table:
    print(record)

# Access deleted records
for record in table.deleted:
    print(record)

# Get total record count
print(len(table))

# Using context manager
with DBF('people.dbf') as table:
    for record in table:
        print(record)

As of version 2.0, dbfread.open() and dbfread.read() are deprecated and will be removed in version 2.2. You should migrate to using the DBF class directly.

Note that the DBF class is no longer a subclass of list. While iteration and counting still work, you must use the .records attribute for list-like operations such as indexing.

# Old way (deprecated):
# table = dbfread.read('people.dbf')
# print(table[1])

# New way:
from dbfread import DBF
table = DBF('people.dbf', load=True)
print(table.records[1])