SDF to CSV Converter

Convert one or more SDF files into CSV tables using RDKit.

The script extracts molecule properties from SDF records, adds a molecule index, computes canonical SMILES, and attempts to generate an InChIKey for each readable molecule.

Features

• Accepts individual SDF files, directories, or glob patterns.
• Writes one CSV per SDF by default.
• Can combine multiple SDF files into one CSV.
• Recursively searches directories for .sdf files.
• Attempts to continue when individual molecules cannot be sanitized.
• Adds mol_index, Name, SMILES, and InChIKey columns.

Requirements

• Python 3.8+
• pandas
• RDKit

Recommended installation with conda:

conda install -c conda-forge rdkit pandas

Basic Usage

Convert a single SDF file:

python sdf2csv.py compounds.sdf

This writes:

compounds.csv

in the current directory.

Output Directory

Write the CSV file to a specific folder:

python sdf2csv.py compounds.sdf --outdir csv_out

or:

python sdf2csv.py compounds.sdf -o csv_out

Convert Multiple Files

Convert all SDF files in a folder:

python sdf2csv.py sdf_files --outdir csv_out

Convert files matching a pattern:

python sdf2csv.py *.sdf --outdir csv_out

Combine Outputs

Combine all input SDF files into a single CSV:

python sdf2csv.py sdf_files --outdir csv_out --combine

Set the combined output filename:

python sdf2csv.py sdf_files --outdir csv_out --combine --outname all_compounds.csv

When --combine is used, the output includes a source_file column.

Verbose Mode

Print progress and warnings:

python sdf2csv.py sdf_files --outdir csv_out --verbose

Output Columns

The script prioritizes these columns:

• mol_index
• Name
• SMILES
• InChIKey

Any additional SDF properties are preserved as extra CSV columns.

Notes

Unreadable molecules are skipped with warnings. If a molecule cannot be sanitized, the script still tries to extract available properties and compute SMILES/InChIKey.