python-xlsxio

Wrapper for c library xlsxio

Since the main library is written in c, and its wrapper is written in cython (which compiles in c), the read speed is much faster than libraries written in python (as can be seen in the benchmarks).

This library also has streaming reading and does not load memory.

Of the minuses, you should know in advance the data types of each column, or read everything as a string or bytes.

Install

Just: pip install python-xlsxio

If you want speed up about 5% you can build from source:

pip install cython python-xlsxio --no-binary python-xlsxio

Benchmarks read xlsx

How is testing you can see in benchmark/benchmark.py

File xlsx generated with openpyxl and conteins mixed types, how is generated you can see in benchmark/generate_xlsx.py

Big file - file with 100000 rows

Small file - file with 100 rows

Test machine: Linux 5.4.44-1-MANJARO #1 SMP PREEMPT Wed Jun 3 14:48:07 UTC 2020 x86_64

Title	Min sec execution	Calls/sec
xlsxio read big file with types	1.08584	0.92095
xlsxio read big file without types	0.93787	1.06625
xlsxio read big file without types all in bytes	0.88268	1.13292
xlsxio read big file with types from memory	1.09569	0.91267
xlsxio read small file with types	0.00128	780.58583
xlsxio read small file without types	0.00117	855.33007
xlsxio read small file without types all in bytes	0.0011	905.32902
xlsxio read small file with types from memory	0.00129	776.06377
openpyxl read big file	6.34512	0.1576
openpyxl read big file from memory	6.35287	0.15741
openpyxl read small file	0.01131	88.40829
openpyxl read small file from memory	0.01134	88.19221
xlrd read big file	4.10823	0.24341
xlrd read big file from memory	4.0911	0.24443
xlrd read small file	0.00486	205.74491
xlrd read small file from memory	0.00503	198.75329
sxl read big file	4.9788	0.20085
sxl read big file from memory	4.96794	0.20129
sxl read small file	0.00627	159.49498
sxl read small file from memory	0.00619	161.4445
xlsx2csv read big file	6.53466	0.15303
xlsx2csv read big file from memory	6.56024	0.15243
xlsx2csv read small file	0.01157	86.4205
xlsx2csv read small file from memory	0.01147	87.21896

Fast start with read xlsx

import xlsxio
xlsxio_reader = xlsxio.XlsxioReader('file.xlsx')
sheet = xlsxio_reader.get_sheet()
data = sheet.read_data()
sheet.close()
xlsxio_reader.close()

print(data)

Or simply:

import xlsxio
with xlsxio.XlsxioReader('file.xlsx') as reader:
    with reader.get_sheet() as sheet:
        data = sheet.read_data()

print(data)

Full example for reading xlsx file in sheet hello, and not reading at all in memory (write only rows, which have True in 5 column):

import xlsxio
import datetime

types = [str, str, float, int, bool, datetime.datetime]
with xlsxio.XlsxioReader('file.xlsx') as reader:
    with reader.get_sheet('hello', types=types) as sheet:
        header = sheet.read_header()
        only_active = []
        for row in sheet.iter_rows():
            if row[4]:
                only_active.append(row)
print(only_active)

Usage read xlsx

XlsxioReader

Object of xlsx

def __init__(self, filename, encoding: str = 'utf-8')

Inittializating XlsxioReader

• filename - str (path to filename), bytes (loaded in memory file) or file like object (not BytesIO)
• encoding - encoding of xlsx file

def get_sheet_names(self) -> tuple

Return tuple of sheet names in xlsx file

def get_sheet(self, sheetname: Optional[str] = None, flags: int = XlsxioReadFlag.SKIP_EMPTY_ROWS, types: Optional[Iterable[type]] = None, default_type: type = str) -> XlsxioReaderSheet

Return XlsxioReaderSheet object

• sheetname - name of sheet (if None, returns first sheet)
• flags - default is XlsxioReadFlag.SKIP_EMPTY_ROWS (Read more about flags). All possible flags:
  • XlsxioReadFlag.SKIP_NONE
  • XlsxioReadFlag.SKIP_EMPTY_ROWS
  • XlsxioReadFlag.SKIP_EMPTY_CELLS
  • XlsxioReadFlag.SKIP_ALL_EMPTY
  • XlsxioReadFlag.SKIP_EXTRA_CELLS
  • XlsxioReadFlag.SKIP_HIDDEN_ROWS
• types - list of types by columns. example, if first column is integer, second - str, end third - float, you can pass: types=[int, str, float]. if fourth column will be, then will it default_type. Possible types:
  • bytes
  • str
  • int
  • float
  • datetime.datetime
  • bool
• default_type - default type of columns if types not passed, default str

def close(self)

Closes reader

XlsxioReaderSheet

Object of sheet

def __init__(self, xlsxioreader: XlsxioReader, sheetname: Optional[str] = None, flags: int = XlsxioReadFlag.SKIP_EMPTY_ROWS, types: Optional[Iterable[type]] = None, default_type: type = str)

Initializet XlsxioReaderSheet object (it object initializes in xlsxioreader.get_sheet and about params you can read there)

def read_row(self, ignore_type: bool = False) -> Optional[list]

Reading next row in list. If rows does not exists return None

• ignore_type - if this true, return row in default_type (convenient for heading)

def read_header(self) -> Optional[list]

Alias for read_row(True)

def iter_rows(self) -> Iterable[list]

Iterate rows while rows exists

def get_last_row_index(self) -> int:

Getting last row index (returns 0 if not readed yet)

def get_flags(self) -> int:

Getting applied flags

def read_data(self) -> List[list]

Read all sheet rows, and first row in default_type. Method code:

header = self.read_header()
if header is None:
    return []
rows = list(self.iter_rows())
rows.insert(0, header)
return rows

def close(self)

Closes sheet