The dataset viewer is not available because its heuristics could not detect any supported data files. You can try uploading some data files, or configuring the data files location manually.
YAML Metadata Warning:empty or missing yaml metadata in repo card
Check out the documentation for more information.
PubChem Compound Alias Master
Standalone pipeline for building a complete, searchable mapping of PubChem compound CIDs, PubChem page titles, and aliases.
It uses PubChem's official bulk files rather than issuing one API request per compound:
CID-Title.gz: the title shown on each PubChem compound summary page.CID-Synonym-filtered.gz: all structure-consistent names associated with each CID. This is the default.CID-Synonym-unfiltered.gz: every depositor-supplied name, including names PubChem considers inconsistent with the structure.
The bulk inputs are several gigabytes compressed. The resulting SQLite database and optional exports require substantially more disk space.
Quick verification
Query PubChem directly without downloading the bulk files:
python pubchem_alias_master/build_pubchem_alias_master.py lookup ceritinib --online
The result includes aliases such as ceritinib, 1032900-25-6, LDK378,
ZYKADIA, and LDK-378.
Build the complete master database
python pubchem_alias_master/build_pubchem_alias_master.py all
Downloads are resumable and verified against PubChem's official MD5 files. Imports checkpoint at complete CID boundaries and can be resumed by rerunning the same command.
For every depositor-supplied alias, including potentially inconsistent names:
python pubchem_alias_master/build_pubchem_alias_master.py all \
--synonym-set unfiltered \
--database pubchem_alias_master/pubchem_aliases_unfiltered.sqlite
Query the local database
python pubchem_alias_master/build_pubchem_alias_master.py lookup ceritinib
python pubchem_alias_master/build_pubchem_alias_master.py lookup 57379345
python pubchem_alias_master/build_pubchem_alias_master.py stats
Export
One JSON object per compound, with aliases kept as a JSON array:
python pubchem_alias_master/build_pubchem_alias_master.py export \
--format jsonl --output pubchem_alias_master/all_compounds.jsonl
One row per CID-alias relationship:
python pubchem_alias_master/build_pubchem_alias_master.py export \
--format edges-tsv --output pubchem_alias_master/all_alias_edges.tsv
One CSV row per compound, with aliases encoded as a lossless JSON array inside
one CSV field. A .gz output suffix enables gzip compression automatically:
python pubchem_alias_master/build_pubchem_alias_master.py export \
--format compounds-csv --output pubchem_alias_master/all_compounds.csv.gz
Export only Ceritinib as a small example:
python pubchem_alias_master/build_pubchem_alias_master.py export \
--format compounds-csv --cid 57379345 \
--output pubchem_alias_master/ceritinib.csv
The complete CSV contains more than 123 million rows and will be very large. The compressed CSV is recommended for archival or transfer; SQLite remains better for interactive searching.
SQLite is the recommended master format because it preserves aliases without CSV delimiter ambiguity and supports exact case-insensitive alias lookup.
Database schema
compounds(cid, title): one row per PubChem compound.aliases(cid, alias_rank, alias): all aliases in PubChem's best-name-first order.metadata(key, value): source and resume information.
Sources:
- Downloads last month
- 105