Quarterly and annual statements (ITR & DFP)

CVM publishes two financial-statement datasets with the same eleven tables and the same schema, differing only in cadence:

ITR — Informações Trimestrais. One CSV per table per year, with one row per (company, quarter, account).
DFP — Demonstrações Financeiras Padronizadas. One CSV per table per year, with one row per (company, fiscal year, account).

The package’s tooling is identical for both: pass "itr" or "dfp" as dataset, the same table name, and the same arguments. Eight tables out of eleven publish individual / consolidated variants selected through report_type.

library(cvmdata)

Reference: tables available

Table	What it carries	`report_type` required?	Notes
`bpa`	Asset side of the balance sheet	Yes (`ind` / `con`)	`VL_CONTA × ESCALA_MOEDA` applied (`multiply_by_scale`).
`bpp`	Liabilities + equity side	Yes	Idem.
`dre`	Income statement	Yes	Idem.
`dra`	Statement of comprehensive income	Yes	Idem.
`dfc_md`	Cash-flow statement, direct method	Yes	Idem.
`dfc_mi`	Cash-flow statement, indirect method	Yes	Idem.
`dmpl`	Statement of changes in equity	Yes	Idem; wider table (multiple equity columns).
`dva`	Statement of value added (Brazilian-specific)	Yes	Idem.
`composicao_capital`	Capital composition (share counts by class)	No (`NULL`)	No `cd_cvm` column — CD_CVM filters resolve via `submissao`.
`parecer`	Auditor opinion (free-text)	No	One row per filing, with full opinion text in `texto`.
`submissao`	Filing header — `id_doc`, `dt_receb`, `link_doc`, etc.	No	The dataset’s index. Every filing has exactly one row here.

Use cvm_dictionary("dfp", "<table>") to see the column-level metadata (descriptions, domain, data type, size) for any of the eleven tables; cvm_codelist("dfp", "<table>", "<column>") returns enumerated categorical values.

Workflow 1 — Side-by-side total assets across two issuers

Goal: compare 2024 total assets for BCO BRASIL S.A. (CD_CVM 1023) and MAGAZINE LUIZA S.A. (CD_CVM 22470) using the individual balance sheet (dfp/bpa).

bpa <- issuer_fetch(
  "dfp", "bpa",
  report_type = "ind",
  issuer = c("1023", "22470"),
  year = 2024
)
#> duckdb keeps downloaded extensions and secrets in a temporary directory:
#> ℹ /tmp/RtmpeyV2ID/duckdb
#> This is removed when the R session ends.
#> • Extensions are re-downloaded each session.
#> • Secrets are lost.
#> ℹ Run duckdb(shared_home = TRUE) (or create ~/.duckdb) to keep them (suitable for most users).
#> ℹ Run duckdb(shared_home = FALSE) to accept the temporary directory (and silence this message).
#> ℹ See ?duckdb_storage for details and alternatives.
bpa
#> ℹ source: "mirror" | fetched_at: 2026-08-25 23:59:37.511781
#> ℹ group: "companhias" | dataset: "dfp" | table: "bpa"
#> # A tibble: 226 × 15
#>    cnpj_cia       dt_refer   versao denom_cia cd_cvm grupo_dfp moeda ordem_exerc
#>    <chr>          <date>     <chr>  <chr>     <chr>  <chr>     <chr> <chr>      
#>  1 00.000.000/00… 2024-12-31 1      BCO BRAS… 001023 DF Indiv… REAL  PENÚLTIMO  
#>  2 00.000.000/00… 2024-12-31 1      BCO BRAS… 001023 DF Indiv… REAL  ÚLTIMO     
#>  3 00.000.000/00… 2024-12-31 1      BCO BRAS… 001023 DF Indiv… REAL  PENÚLTIMO  
#>  4 00.000.000/00… 2024-12-31 1      BCO BRAS… 001023 DF Indiv… REAL  ÚLTIMO     
#>  5 00.000.000/00… 2024-12-31 1      BCO BRAS… 001023 DF Indiv… REAL  PENÚLTIMO  
#>  6 00.000.000/00… 2024-12-31 1      BCO BRAS… 001023 DF Indiv… REAL  ÚLTIMO     
#>  7 00.000.000/00… 2024-12-31 1      BCO BRAS… 001023 DF Indiv… REAL  PENÚLTIMO  
#>  8 00.000.000/00… 2024-12-31 1      BCO BRAS… 001023 DF Indiv… REAL  ÚLTIMO     
#>  9 00.000.000/00… 2024-12-31 1      BCO BRAS… 001023 DF Indiv… REAL  PENÚLTIMO  
#> 10 00.000.000/00… 2024-12-31 1      BCO BRAS… 001023 DF Indiv… REAL  ÚLTIMO     
#> # ℹ 216 more rows
#> # ℹ 7 more variables: dt_fim_exerc <date>, cd_conta <chr>, ds_conta <chr>,
#> #   vl_conta <dbl>, st_conta_fixa <chr>, report_type <chr>, year <int>

The tibble carries provenance attributes and the package has already:

multiplied VL_CONTA by ESCALA_MOEDA and dropped the scale column (values are in absolute reais);
converted DT_REFER, DT_FIM_EXERC to R Date;
kept only the latest VERSAO per (cnpj_cia, dt_refer) — so re-filings do not produce duplicated rows.

Total assets is the top of the asset hierarchy (cd_conta == "1"):

ativo_total <- bpa[
  bpa$cd_conta == "1" & bpa$ordem_exerc == "ÚLTIMO",
  c("cnpj_cia", "denom_cia", "dt_fim_exerc", "vl_conta")
]
ativo_total
#> ℹ source: "mirror" | fetched_at: 2026-08-25 23:59:37.511781
#> ℹ group: "companhias" | dataset: "dfp" | table: "bpa"
#> # A tibble: 2 × 4
#>   cnpj_cia           denom_cia           dt_fim_exerc      vl_conta
#>   <chr>              <chr>               <date>               <dbl>
#> 1 00.000.000/0001-91 BCO BRASIL S.A.     2024-12-31   2395432208000
#> 2 47.960.950/0001-21 MAGAZINE LUIZA S.A. 2024-12-31     32482619000

CVM publishes each fiscal year’s CSV with two reference columns — ÚLTIMO (the declared year, here 2024) and PENÚLTIMO (the comparative previous year). To pick up an older comparative, request the matching upstream year explicitly:

bpa_history <- issuer_fetch(
  "dfp", "bpa",
  report_type = "ind",
  issuer = "1023",
  year = 2022:2024
)
#> duckdb keeps downloaded extensions and secrets in a temporary directory:
#> ℹ /tmp/RtmpeyV2ID/duckdb
#> This is removed when the R session ends.
#> • Extensions are re-downloaded each session.
#> • Secrets are lost.
#> ℹ Run duckdb(shared_home = TRUE) (or create ~/.duckdb) to keep them (suitable for most users).
#> ℹ Run duckdb(shared_home = FALSE) to accept the temporary directory (and silence this message).
#> ℹ See ?duckdb_storage for details and alternatives.
ativo_history <- bpa_history[
  bpa_history$cd_conta == "1" & bpa_history$ordem_exerc == "ÚLTIMO",
  c("dt_fim_exerc", "vl_conta")
]
ativo_history[order(ativo_history$dt_fim_exerc), ]
#> ℹ source: "mirror" | fetched_at: 2026-08-25 23:59:38.536401
#> ℹ group: "companhias" | dataset: "dfp" | table: "bpa"
#> # A tibble: 3 × 2
#>   dt_fim_exerc      vl_conta
#>   <date>               <dbl>
#> 1 2022-12-31   2062674549000
#> 2 2023-12-31   2208053634000
#> 3 2024-12-31   2395432208000

Workflow 2 — Capital composition through `submissao` lookup

composicao_capital does not carry cd_cvm — it indexes by cnpj_cia. The package reads the dataset’s submissao table for the same year, looks up CNPJ for each CD_CVM passed in companies, and applies the filter on CNPJ. The user-facing interface is identical:

cap <- issuer_fetch(
  "dfp", "composicao_capital",
  issuer = c("1023", "22470"),
  year = 2024
)
#> duckdb keeps downloaded extensions and secrets in a temporary directory:
#> ℹ /tmp/RtmpeyV2ID/duckdb
#> This is removed when the R session ends.
#> • Extensions are re-downloaded each session.
#> • Secrets are lost.
#> ℹ Run duckdb(shared_home = TRUE) (or create ~/.duckdb) to keep them (suitable for most users).
#> ℹ Run duckdb(shared_home = FALSE) to accept the temporary directory (and silence this message).
#> ℹ See ?duckdb_storage for details and alternatives.
#> ℹ Resolving CD_CVM 1023, 22470 via "dfp"/submissao for 2024 (table
#>   "composicao_capital" does not carry `cd_cvm`).
cap
#> ℹ source: "mirror" | fetched_at: 2026-08-25 23:59:38.963918
#> ℹ group: "companhias" | dataset: "dfp" | table: "composicao_capital"
#> # A tibble: 2 × 11
#>   cnpj_cia           dt_refer   versao denom_cia          qt_acao_ordin_cap_in…¹
#>   <chr>              <date>     <chr>  <chr>                               <dbl>
#> 1 00.000.000/0001-91 2024-12-31 1      BCO BRASIL S.A.                5730834040
#> 2 47.960.950/0001-21 2024-12-31 1      MAGAZINE LUIZA S.…              738995248
#> # ℹ abbreviated name: ¹qt_acao_ordin_cap_integr
#> # ℹ 6 more variables: qt_acao_pref_cap_integr <dbl>,
#> #   qt_acao_total_cap_integr <dbl>, qt_acao_ordin_tesouro <dbl>,
#> #   qt_acao_pref_tesouro <dbl>, qt_acao_total_tesouro <dbl>, year <int>

Note: no report_type argument — composicao_capital is a single concept, not split into individual / consolidated.

Cross-check against the underlying submissao header:

sub <- issuer_fetch(
  "dfp", "submissao",
  issuer = c("1023", "22470"),
  year = 2024
)
#> duckdb keeps downloaded extensions and secrets in a temporary directory:
#> ℹ /tmp/RtmpeyV2ID/duckdb
#> This is removed when the R session ends.
#> • Extensions are re-downloaded each session.
#> • Secrets are lost.
#> ℹ Run duckdb(shared_home = TRUE) (or create ~/.duckdb) to keep them (suitable for most users).
#> ℹ Run duckdb(shared_home = FALSE) to accept the temporary directory (and silence this message).
#> ℹ See ?duckdb_storage for details and alternatives.
sub[, c("cd_cvm", "denom_cia", "dt_refer", "id_doc")]
#> ℹ source: "mirror" | fetched_at: 2026-08-25 23:59:39.390826
#> ℹ group: "companhias" | dataset: "dfp" | table: "submissao"
#> # A tibble: 2 × 4
#>   cd_cvm denom_cia           dt_refer   id_doc
#>   <chr>  <chr>               <date>     <chr> 
#> 1 001023 BCO BRASIL S.A.     2024-12-31 144874
#> 2 022470 MAGAZINE LUIZA S.A. 2024-12-31 145377

ITR works the same way — replace "dfp" with "itr". The trimester breakdown then shows up in dt_fim_exerc and in the ds_conta / grupo_dfp columns; the underlying mechanics (multiply_by_scale, keep_latest_version, date conversion via the dictionary) are shared.

Where to read next

issuer-fetch — argument reference and the full companies / years / report_type / validate / on_error semantics.
fre — reference form. Same package, different table semantics.
cvm-defects — known publication quirks behind the transformations applied here.

Reference: tables available

Workflow 1 — Side-by-side total assets across two issuers

Workflow 2 — Capital composition through submissao lookup

Where to read next

Workflow 2 — Capital composition through `submissao` lookup