Databolt File Ingest

Quickly ingest raw files. Works for XLS, CSV, TXT which can be exported to CSV, Parquet, SQL and Pandas. d6tstack solves many performance and schema problems typically encountered when ingesting raw files.

Features include

Fast pd.to_sql() for postgres and mysql
Quickly check columns for consistency across files
Fix added/missing columns
Fix renamed columns
Check Excel tabs for consistency across files
Excel to CSV converter (incl multi-sheet support)
Out of core functionality to process large files
Export to CSV, parquet, SQL, pandas dataframe

Sample Use


import d6tstack

# fast CSV to SQL import - see SQL examples notebook
d6tstack.utils.pd_to_psql(df, 'postgresql+psycopg2://usr:pwd@localhost/db', 'tablename')
d6tstack.utils.pd_to_mysql(df, 'mysql+mysqlconnector://usr:pwd@localhost/db', 'tablename')
d6tstack.utils.pd_to_mssql(df, 'mssql+pymssql://usr:pwd@localhost/db', 'tablename') # experimental

# ingest mutiple CSVs which may have data schema changes - see CSV examples notebook

import glob
>>> c = d6tstack.combine_csv.CombinerCSV(glob.glob('data/*.csv'))

# quick check if all files have consistent columns
>>> c.is_all_equal()
False

# show which files have missing columns
>>> c.is_col_present()
   filename  cost  date profit profit2 sales
0  feb.csv  True  True   True   False  True
2  mar.csv  True  True   True    True  True

>>> c.combine_preview() # keep all columns
   filename  cost        date profit profit2 sales
0   jan.csv  -80  2011-01-01     20     NaN   100
0   mar.csv  -100  2011-03-01    200     400   300

>>> d6tstack.combine_csv.CombinerCSV(glob.glob('*.csv'), columns_select_common=True).combine_preview() # keep common columns
   filename  cost        date profit sales
0   jan.csv  -80  2011-01-01     20   100
0   mar.csv  -100  2011-03-01    200   300

>>> d6tstack.combine_csv.CombinerCSV(glob.glob('*.csv'), columns_rename={'sales':'revenue'}).combine_preview()
   filename  cost        date profit profit2 revenue
0   jan.csv  -80  2011-01-01     20     NaN   100
0   mar.csv  -100  2011-03-01    200     400   300

# to come: check if columns match database
>>> c.is_columns_match_db('postgresql+psycopg2://usr:pwd@localhost/db', 'tablename')

# export to csv, parquet, sql. Out of core with optimized fast imports for postgres and mysql
>>> c.to_pandas()
>>> c.to_csv_align(output_dir='process/')
>>> c.to_parquet_align(output_dir='process/')
>>> c.to_sql_combine('postgresql+psycopg2://usr:pwd@localhost/db', 'tablename')
>>> c.to_psql_combine('postgresql+psycopg2://usr:pwd@localhost/db', 'tablename') # fast, using COPY FROM
>>> c.to_mysql_combine('mysql+mysqlconnector://usr:pwd@localhost/db', 'tablename') # fast, using LOAD DATA LOCAL INFILE

# read Excel files - see Excel examples notebook for more details
import d6tstack.convert_xls

d6tstack.convert_xls.read_excel_advanced('test.xls',
    sheet_name='Sheet1', header_xls_range="B2:E2")

d6tstack.convert_xls.XLStoCSVMultiSheet('test.xls').convert_all(header_xls_range="B2:E2")

d6tstack.convert_xls.XLStoCSVMultiFile(glob.glob('*.xls'), 
    cfg_xls_sheets_sel_mode='name_global',cfg_xls_sheets_sel='Sheet1')
    .convert_all(header_xls_range="B2:E2")

Installation

We recommend using the latest version from github pip install git+https://github.com/d6t/d6tstack.git.

If you cannot install from github, use the latest published version pip install d6tstack. For Excel and parquet support, install d6tstack[xls] and d6tstack[parquet]. Certain database specific function require packages which you will be prompted for as you use them.

Documentation

CSV examples notebook - Quickly load any type of CSV files
Excel examples notebook - Quickly extract from Excel to CSV
Dask Examples notebook - How to use d6tstack to solve Dask input file problems
Pyspark Examples notebook - How to use d6tstack to solve pyspark input file problems
SQL examples notebook - Fast loading of CSV to SQL with pandas preprocessing
Function reference docs - Detailed documentation for modules, classes, functions
www.databolt.tech - Web app if you don't want to code

Faster Data Engineering

Check out other d6t libraries to solve common data engineering problems, including

data ingest, quickly ingest raw data
fuzzy joins, quickly join data
data pipes, quickly share and distribute data

https://github.com/d6t/d6t-python

And we encourage you to join the Databolt blog to get updates and tips+tricks http://blog.databolt.tech

Collecting Errors Messages and Usage statistics

To help us make this library better, it collects anonymous error messages and usage statistics. It works similar to how websites collect data. See d6tcollect for details including how to disable collection.

It might not catch all errors so if you run into any problems, please raise an issue on github.