Skip to content

Releases

Mirror: companhias/recompra_acoes (snapshot 2026-11-18)

Choose a tag to compare

@github-actions github-actions released this 18 Nov 14:45
53954ae

Dated snapshot of the parquet mirror for group 'companhias', dataset 'recompra_acoes', as of 2026-11-18 (published 2026-11-18 14:45:24 UTC).

Carries the 0 partition(s) whose content changed since the previous snapshot, out of 3 in the mirror. __source_hash.json lists every partition of this snapshot and, under releases, the snapshot release that holds its bytes. Read it from R with options(cvmdata.mirror_release = "2026-11-18"). Snapshot releases are never deleted: later snapshots point into this one.

What is here. One parquet file per table, statement basis and reference year (<table>__report_type=<ind|con>__year=<YYYY>__part-0.parquet; GitHub shows = as .), written by the ETL of the package (inst/etl/ in the repository) from the CSV files the CVM publishes for the dataset recompra_acoes: column names in lower case, dates and numbers parsed, values as the CVM delivers them (no row dropped, no value corrected). The __history.parquet assets keep the superseded versions of each partition (restated filings), appended week after week. Column dictionary: cvm_dictionary() in the package or https://sidneybissoli.github.io/cvmdata/reference/cvm_dictionary.html; code lists: cvm_codelist(). __source_hash.json is the index of the release (content hash, digest of every partition, snapshot dates). Quarterly editions of the whole mirror, with a DOI, are planned on Zenodo from January 2027.

Data licence. Contains data from the Comissao de Valores Mobiliarios (CVM) Open Data Portal, https://dados.cvm.gov.br/, made available under the Open Data Commons Open Database License (ODbL) v1.0, https://opendatacommons.org/licenses/odbl/1-0/. The parquet files in this release are a derivative database of that portal, converted by the cvmdata ETL, and are redistributed under the same licence. Attribution and terms: https://github.com/SidneyBissoli/cvmdata/blob/main/DATA-LICENSE.md

Personal data. The mirror holds data about companies, not about people. What it leaves out, who answers for it and how a person mentioned in a text can ask for the name to be removed: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.md (em portugues: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.pt-BR.md)

Mirror: companhias/ipe (snapshot 2026-11-18)

Choose a tag to compare

@github-actions github-actions released this 18 Nov 14:45
53954ae

Dated snapshot of the parquet mirror for group 'companhias', dataset 'ipe', as of 2026-11-18 (published 2026-11-18 14:45:31 UTC).

Carries the 0 partition(s) whose content changed since the previous snapshot, out of 24 in the mirror. __source_hash.json lists every partition of this snapshot and, under releases, the snapshot release that holds its bytes. Read it from R with options(cvmdata.mirror_release = "2026-11-18"). Snapshot releases are never deleted: later snapshots point into this one.

What is here. One parquet file per table, statement basis and reference year (<table>__report_type=<ind|con>__year=<YYYY>__part-0.parquet; GitHub shows = as .), written by the ETL of the package (inst/etl/ in the repository) from the CSV files the CVM publishes for the dataset ipe: column names in lower case, dates and numbers parsed, values as the CVM delivers them (no row dropped, no value corrected). The __history.parquet assets keep the superseded versions of each partition (restated filings), appended week after week. Column dictionary: cvm_dictionary() in the package or https://sidneybissoli.github.io/cvmdata/reference/cvm_dictionary.html; code lists: cvm_codelist(). __source_hash.json is the index of the release (content hash, digest of every partition, snapshot dates). Quarterly editions of the whole mirror, with a DOI, are planned on Zenodo from January 2027.

Data licence. Contains data from the Comissao de Valores Mobiliarios (CVM) Open Data Portal, https://dados.cvm.gov.br/, made available under the Open Data Commons Open Database License (ODbL) v1.0, https://opendatacommons.org/licenses/odbl/1-0/. The parquet files in this release are a derivative database of that portal, converted by the cvmdata ETL, and are redistributed under the same licence. Attribution and terms: https://github.com/SidneyBissoli/cvmdata/blob/main/DATA-LICENSE.md

Personal data. The mirror holds data about companies, not about people. What it leaves out, who answers for it and how a person mentioned in a text can ask for the name to be removed: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.md (em portugues: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.pt-BR.md)

Mirror: companhias/cad (snapshot 2026-11-18)

Choose a tag to compare

@github-actions github-actions released this 18 Nov 14:45
53954ae

Dated snapshot of the parquet mirror for group 'companhias', dataset 'cad', as of 2026-11-18 (published 2026-11-18 14:45:29 UTC).

Carries the 1 partition(s) whose content changed since the previous snapshot, out of 3 in the mirror. __source_hash.json lists every partition of this snapshot and, under releases, the snapshot release that holds its bytes. Read it from R with options(cvmdata.mirror_release = "2026-11-18"). Snapshot releases are never deleted: later snapshots point into this one.

What is here. One parquet file per table, statement basis and reference year (<table>__report_type=<ind|con>__year=<YYYY>__part-0.parquet; GitHub shows = as .), written by the ETL of the package (inst/etl/ in the repository) from the CSV files the CVM publishes for the dataset cad: column names in lower case, dates and numbers parsed, values as the CVM delivers them (no row dropped, no value corrected). The __history.parquet assets keep the superseded versions of each partition (restated filings), appended week after week. Column dictionary: cvm_dictionary() in the package or https://sidneybissoli.github.io/cvmdata/reference/cvm_dictionary.html; code lists: cvm_codelist(). __source_hash.json is the index of the release (content hash, digest of every partition, snapshot dates). Quarterly editions of the whole mirror, with a DOI, are planned on Zenodo from January 2027.

Data licence. Contains data from the Comissao de Valores Mobiliarios (CVM) Open Data Portal, https://dados.cvm.gov.br/, made available under the Open Data Commons Open Database License (ODbL) v1.0, https://opendatacommons.org/licenses/odbl/1-0/. The parquet files in this release are a derivative database of that portal, converted by the cvmdata ETL, and are redistributed under the same licence. Attribution and terms: https://github.com/SidneyBissoli/cvmdata/blob/main/DATA-LICENSE.md

Personal data. The mirror holds data about companies, not about people. What it leaves out, who answers for it and how a person mentioned in a text can ask for the name to be removed: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.md (em portugues: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.pt-BR.md)

Mirror: companhias/recompra_acoes (snapshot 2026-11-17)

Choose a tag to compare

@github-actions github-actions released this 17 Nov 16:39
eae8d74

Dated snapshot of the parquet mirror for group 'companhias', dataset 'recompra_acoes', as of 2026-11-17 (published 2026-11-17 17:04:38 UTC).

Carries the 3 partition(s) whose content changed since the previous snapshot, out of 3 in the mirror. __source_hash.json lists every partition of this snapshot and, under releases, the snapshot release that holds its bytes. Read it from R with options(cvmdata.mirror_release = "2026-11-17"). Snapshot releases are never deleted: later snapshots point into this one.

What is here. One parquet file per table, statement basis and reference year (<table>__report_type=<ind|con>__year=<YYYY>__part-0.parquet; GitHub shows = as .), written by the ETL of the package (inst/etl/ in the repository) from the CSV files the CVM publishes for the dataset recompra_acoes: column names in lower case, dates and numbers parsed, values as the CVM delivers them (no row dropped, no value corrected). The __history.parquet assets keep the superseded versions of each partition (restated filings), appended week after week. Column dictionary: cvm_dictionary() in the package or https://sidneybissoli.github.io/cvmdata/reference/cvm_dictionary.html; code lists: cvm_codelist(). __source_hash.json is the index of the release (content hash, digest of every partition, snapshot dates). Quarterly editions of the whole mirror, with a DOI, are planned on Zenodo from January 2027.

Data licence. Contains data from the Comissao de Valores Mobiliarios (CVM) Open Data Portal, https://dados.cvm.gov.br/, made available under the Open Data Commons Open Database License (ODbL) v1.0, https://opendatacommons.org/licenses/odbl/1-0/. The parquet files in this release are a derivative database of that portal, converted by the cvmdata ETL, and are redistributed under the same licence. Attribution and terms: https://github.com/SidneyBissoli/cvmdata/blob/main/DATA-LICENSE.md

Personal data. The mirror holds data about companies, not about people. What it leaves out, who answers for it and how a person mentioned in a text can ask for the name to be removed: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.md (em portugues: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.pt-BR.md)

Mirror: companhias/itr (snapshot 2026-11-17)

Choose a tag to compare

@github-actions github-actions released this 17 Nov 16:52
eae8d74

Dated snapshot of the parquet mirror for group 'companhias', dataset 'itr', as of 2026-11-17 (published 2026-11-17 16:52:29 UTC).

Carries the 0 partition(s) whose content changed since the previous snapshot, out of 295 in the mirror. __source_hash.json lists every partition of this snapshot and, under releases, the snapshot release that holds its bytes. Read it from R with options(cvmdata.mirror_release = "2026-11-17"). Snapshot releases are never deleted: later snapshots point into this one.

What is here. One parquet file per table, statement basis and reference year (<table>__report_type=<ind|con>__year=<YYYY>__part-0.parquet; GitHub shows = as .), written by the ETL of the package (inst/etl/ in the repository) from the CSV files the CVM publishes for the dataset itr: column names in lower case, dates and numbers parsed, values as the CVM delivers them (no row dropped, no value corrected). The __history.parquet assets keep the superseded versions of each partition (restated filings), appended week after week. Column dictionary: cvm_dictionary() in the package or https://sidneybissoli.github.io/cvmdata/reference/cvm_dictionary.html; code lists: cvm_codelist(). __source_hash.json is the index of the release (content hash, digest of every partition, snapshot dates). Quarterly editions of the whole mirror, with a DOI, are planned on Zenodo from January 2027.

Data licence. Contains data from the Comissao de Valores Mobiliarios (CVM) Open Data Portal, https://dados.cvm.gov.br/, made available under the Open Data Commons Open Database License (ODbL) v1.0, https://opendatacommons.org/licenses/odbl/1-0/. The parquet files in this release are a derivative database of that portal, converted by the cvmdata ETL, and are redistributed under the same licence. Attribution and terms: https://github.com/SidneyBissoli/cvmdata/blob/main/DATA-LICENSE.md

Personal data. The mirror holds data about companies, not about people. What it leaves out, who answers for it and how a person mentioned in a text can ask for the name to be removed: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.md (em portugues: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.pt-BR.md)

Mirror: companhias/fre (snapshot 2026-11-17)

Choose a tag to compare

@github-actions github-actions released this 17 Nov 16:42
eae8d74

Dated snapshot of the parquet mirror for group 'companhias', dataset 'fre', as of 2026-11-17 (published 2026-11-17 16:42:28 UTC).

Carries the 0 partition(s) whose content changed since the previous snapshot, out of 578 in the mirror. __source_hash.json lists every partition of this snapshot and, under releases, the snapshot release that holds its bytes. Read it from R with options(cvmdata.mirror_release = "2026-11-17"). Snapshot releases are never deleted: later snapshots point into this one.

What is here. One parquet file per table, statement basis and reference year (<table>__report_type=<ind|con>__year=<YYYY>__part-0.parquet; GitHub shows = as .), written by the ETL of the package (inst/etl/ in the repository) from the CSV files the CVM publishes for the dataset fre: column names in lower case, dates and numbers parsed, values as the CVM delivers them (no row dropped, no value corrected). The __history.parquet assets keep the superseded versions of each partition (restated filings), appended week after week. Column dictionary: cvm_dictionary() in the package or https://sidneybissoli.github.io/cvmdata/reference/cvm_dictionary.html; code lists: cvm_codelist(). __source_hash.json is the index of the release (content hash, digest of every partition, snapshot dates). Quarterly editions of the whole mirror, with a DOI, are planned on Zenodo from January 2027.

Data licence. Contains data from the Comissao de Valores Mobiliarios (CVM) Open Data Portal, https://dados.cvm.gov.br/, made available under the Open Data Commons Open Database License (ODbL) v1.0, https://opendatacommons.org/licenses/odbl/1-0/. The parquet files in this release are a derivative database of that portal, converted by the cvmdata ETL, and are redistributed under the same licence. Attribution and terms: https://github.com/SidneyBissoli/cvmdata/blob/main/DATA-LICENSE.md

Personal data. The mirror holds data about companies, not about people. What it leaves out, who answers for it and how a person mentioned in a text can ask for the name to be removed: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.md (em portugues: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.pt-BR.md)

Mirror: companhias/cgvn (snapshot 2026-11-17)

Choose a tag to compare

@github-actions github-actions released this 17 Nov 16:39
eae8d74

Dated snapshot of the parquet mirror for group 'companhias', dataset 'cgvn', as of 2026-11-17 (published 2026-11-17 16:39:52 UTC).

Carries the 0 partition(s) whose content changed since the previous snapshot, out of 18 in the mirror. __source_hash.json lists every partition of this snapshot and, under releases, the snapshot release that holds its bytes. Read it from R with options(cvmdata.mirror_release = "2026-11-17"). Snapshot releases are never deleted: later snapshots point into this one.

What is here. One parquet file per table, statement basis and reference year (<table>__report_type=<ind|con>__year=<YYYY>__part-0.parquet; GitHub shows = as .), written by the ETL of the package (inst/etl/ in the repository) from the CSV files the CVM publishes for the dataset cgvn: column names in lower case, dates and numbers parsed, values as the CVM delivers them (no row dropped, no value corrected). The __history.parquet assets keep the superseded versions of each partition (restated filings), appended week after week. Column dictionary: cvm_dictionary() in the package or https://sidneybissoli.github.io/cvmdata/reference/cvm_dictionary.html; code lists: cvm_codelist(). __source_hash.json is the index of the release (content hash, digest of every partition, snapshot dates). Quarterly editions of the whole mirror, with a DOI, are planned on Zenodo from January 2027.

Data licence. Contains data from the Comissao de Valores Mobiliarios (CVM) Open Data Portal, https://dados.cvm.gov.br/, made available under the Open Data Commons Open Database License (ODbL) v1.0, https://opendatacommons.org/licenses/odbl/1-0/. The parquet files in this release are a derivative database of that portal, converted by the cvmdata ETL, and are redistributed under the same licence. Attribution and terms: https://github.com/SidneyBissoli/cvmdata/blob/main/DATA-LICENSE.md

Personal data. The mirror holds data about companies, not about people. What it leaves out, who answers for it and how a person mentioned in a text can ask for the name to be removed: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.md (em portugues: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.pt-BR.md)

Mirror: companhias/cad (snapshot 2026-11-17)

Choose a tag to compare

@github-actions github-actions released this 17 Nov 16:38
eae8d74

Dated snapshot of the parquet mirror for group 'companhias', dataset 'cad', as of 2026-11-17 (published 2026-11-17 16:38:54 UTC).

Carries the 2 partition(s) whose content changed since the previous snapshot, out of 3 in the mirror. __source_hash.json lists every partition of this snapshot and, under releases, the snapshot release that holds its bytes. Read it from R with options(cvmdata.mirror_release = "2026-11-17"). Snapshot releases are never deleted: later snapshots point into this one.

What is here. One parquet file per table, statement basis and reference year (<table>__report_type=<ind|con>__year=<YYYY>__part-0.parquet; GitHub shows = as .), written by the ETL of the package (inst/etl/ in the repository) from the CSV files the CVM publishes for the dataset cad: column names in lower case, dates and numbers parsed, values as the CVM delivers them (no row dropped, no value corrected). The __history.parquet assets keep the superseded versions of each partition (restated filings), appended week after week. Column dictionary: cvm_dictionary() in the package or https://sidneybissoli.github.io/cvmdata/reference/cvm_dictionary.html; code lists: cvm_codelist(). __source_hash.json is the index of the release (content hash, digest of every partition, snapshot dates). Quarterly editions of the whole mirror, with a DOI, are planned on Zenodo from January 2027.

Data licence. Contains data from the Comissao de Valores Mobiliarios (CVM) Open Data Portal, https://dados.cvm.gov.br/, made available under the Open Data Commons Open Database License (ODbL) v1.0, https://opendatacommons.org/licenses/odbl/1-0/. The parquet files in this release are a derivative database of that portal, converted by the cvmdata ETL, and are redistributed under the same licence. Attribution and terms: https://github.com/SidneyBissoli/cvmdata/blob/main/DATA-LICENSE.md

Personal data. The mirror holds data about companies, not about people. What it leaves out, who answers for it and how a person mentioned in a text can ask for the name to be removed: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.md (em portugues: https://github.com/SidneyBissoli/cvmdata/blob/main/PRIVACY.pt-BR.md)

cvmdata 0.3.1

Choose a tag to compare

@SidneyBissoli SidneyBissoli released this 16 Nov 02:49
a5f7f46

Bug fixes on top of 0.3.0. The mirror now returns what the CVM returns.

  • Same columns from both sources. The mirror added a year column (the partition value of its files) that source = "cvm" does not have. cvm_fetch() and cvm_versions() now drop it.
  • No more silent empty result with year = NULL. With a company that has no filing in the latest yearly file, the mirror returned zero rows without a word, while the CVM path stepped back to the year before (up to three years) with a cvmdata_warn_year_fallback warning. Both sources now share one implementation of the step back.

Found by data-raw/revisao-do-dono.R, the maintainer's end-to-end review script, which reads the same tables from both sources.

Install with pak::pak("SidneyBissoli/cvmdata@v0.3.1").

cvmdata 0.3.0

Choose a tag to compare

@SidneyBissoli SidneyBissoli released this 16 Nov 01:00
70a83a0

The version submitted to rOpenSci.

One function, cvm_fetch(), reads the nine datasets of the CVM group of listed companies (104 tables): the company register, annual and quarterly financial statements, the reference form, the registration form, the corporate governance code report, insider trading positions, share buyback programmes and the index of periodic filings. cvm_versions() returns every version of a filing, with the superseded ones recovered from the CVM filing archive.

What changed since 0.2.0, in short:

  • One fetch function. cvm_fetch() replaces the per-group fetchers; companies are selected by CNPJ or B3 ticker, periods by year or by a window of reference dates; basis, reading and as_of choose the statement basis, the original or restated figures of an exercise, and the registrations in force on a date.
  • Filing versions. The current filing by default; every version with cvm_versions(), with the reason of the resubmission and a label for versions replaced only because they were filed in the wrong unit.
  • The mirror. Default source; dated snapshots that can be pinned with cvm_release_set(); SHA-256 checked on every download; a history of superseded filings.
  • Company data only. The mirror and its Zenodo editions carry data about companies; the tables that describe natural persons are read straight from the CVM, by the same call. See PRIVACY.md and DATA-LICENSE.md.
  • A DOI for the data. First Zenodo edition: 10.5281/zenodo.24260441; 10.5281/zenodo.24260440 resolves to the latest edition.

The full list of changes is in NEWS.md.

Install with pak::pak("SidneyBissoli/cvmdata@v0.3.0").