/vcf2df

VCF to parquet converter

Primary LanguageGoMIT LicenseMIT

vcf2df

Reads sample.vcf.gz and writes sample.parquet

Reserved column names:

  • VARIANTKEY (STRING)

    Variantkey is CHR-POS-REF-ALT for small variants and CHR-POS-END-SVTYPE for structural variants. CHR without chr preffix.

  • CHROM (STRING)

    Chromosome (including chr preffix).

  • POS (INT)

    Position (1-based)

  • REF (STRING)

    Reference allele. Empty for structural variants.

  • ALT (STRING)

    Alternate allele. Empty for structural variants.

  • QUAL (INT)

    Quality score (Integer)

  • PASS (BOOL)

    Boolean describing filter == PASS || filter == "."

  • IS_SV (BOOL)

    Boolean structural variant.",

  • SVTYPE (STRING)

    Structural variant type (ex. DEL, DUP, INV, ...)

  • END (INT)

    End position of variant (1-based).

  • NUMALTS (INT)

    Number of alternate alleles (0, 1, 2)

  • SAMPLE (STRING)

    Sample string.

  • IS_PHASED (BOOL)

    Boolean if variant is phased.

  • PHASE_ID (STRING)

    String identifying variant phase.

  • REF_READS (INT)

    Read depth for ref.

  • ALT_READS (INT)

    Read depth for alt.

Samples:

If vcf has 1+ sample genotypes, the parquet file will contain 1 line per sample with > 0 alleles. If the vcf does not contain samples, all variants will be represented in the parquet file.

Footer:

All columns are described in the parquet metadata (footer) key-values.