You are losing the schema on purpose
The CSV that comes out has no types. Whatever reads it next will have to guess again. If the destination understands Parquet, sending the Parquet file is strictly better.
Convert Parquet to CSV
Parquet is excellent for machines and useless for a quick look. If you have been handed a .parquet file and need it in something you can open, this converter decompresses it, flattens it, and writes a plain CSV — locally, with no Python environment to set up.
- Runs in your browser
- No signup
- No upload
- Free
ParquetCSV
Runs on your deviceCSV
Options
Reading Parquet
Auto-detect reads the first rows and picks the separator that splits them consistently.
Turn off when the file starts straight into data. Columns are then named column_1, column_2, and so on.
Turns "42" into a number and "true" into a boolean. Values with leading zeros stay text so IDs and zip codes survive.
Which worksheet to convert. Filled in once the workbook is read.
Dotted path to the array of records, such as data.items. Leave empty to use the whole document.
Turns {"user":{"id":7}} into a user.id column. Arrays are kept as JSON text.
Pick the encoding the file was written in. Wrong encoding shows up as garbled accents.
Comma-separated words that should become null instead of text.
Writing CSV
Snappy is the default across Spark, DuckDB and pandas. Gzip is smaller but slower to read.
Turns a user.id column back into {"user":{"id":7}}.
Excel needs a byte order mark to open UTF-8 files with accents correctly.
Rows per row group. Larger groups compress better; smaller groups let readers skip more.
Converting…
Ad slot ·
conversion-preroll
This short ad is what keeps the converter free.
Converting Parquet just to look inside it?Open Parquet files directly with ParquetReader — opens Parquet files directly.What happens
How this converter handles Parquet to CSV
- Reads the schema from the file footer and decompresses the row groups, including Snappy, Gzip, Zstd, Brotli and LZ4.
- Flattens nested columns into dotted names, so a struct field becomes a user.email column instead of being dropped.
- Converts Parquet types back to text: timestamps to ISO-8601, booleans to true and false, nulls to empty fields.
- Writes a standard comma-separated file, with an optional BOM so Excel opens accented characters correctly.
Format differences
Parquet vs CSV
Typed, compressed, columnar storage built for analytical scans. Plain text rows that open anywhere, with no types and no schema.
| Parquet | CSV | |
|---|---|---|
| Type information | Full — ints, floats, booleans, timestamps, decimals | None — every value is text |
| Structure | Columnar, with nested type support | Flat rows and columns |
| Compression | Built in (Snappy, Gzip, Zstd) | None |
| Typical file size | Typically 5–15% of the CSV | Baseline (100%) |
| Schema | Stored in the file footer | Header row at best |
| Human-readable | No — binary | Yes, in any text editor |
| Tool support | Spark, DuckDB, pandas, BigQuery, Athena | Universal |
Worth knowing
What to watch out for
Large integers can lose precision downstream
Parquet INT64 values beyond about 9 quadrillion are written to the CSV exactly, but a spreadsheet or JavaScript-based reader may round them on the way back in. Snowflake IDs and similar keys are the usual victims.
The file gets much bigger
Expect the CSV to be roughly 7 to 20 times the size of the Parquet file. That is the compression and columnar layout being undone, and it is worth checking you have the disk space before converting a large file.
When you need this
Common reasons to convert Parquet to CSV
Sending data to someone without a data stack
Finance, ops and legal teams open CSVs. Handing them a Parquet file usually means handing them a support request as well.
Importing into a tool that has no Parquet support
Plenty of CRMs, ad platforms and legacy databases only accept delimited text. CSV remains the reliable lowest common denominator.
A fast sanity check on a pipeline output
Converting a small Parquet output to CSV is a quick way to eyeball whether the columns and row counts look right.
FAQ
Parquet to CSV questions
Do I need Python or pandas to open a Parquet file?
Not for this. The Parquet reader is compiled to WebAssembly and runs inside the page, so the browser does what pandas, PyArrow or DuckDB would normally do locally. Nothing is installed and nothing is uploaded. If you open Parquet files often, ParquetReader skips the conversion entirely and reads them in place.
Which compression codecs are supported?
Snappy, Gzip, Zstd, Brotli, LZ4 and uncompressed files all read correctly. Encrypted Parquet files do not, since they require a key exchange this tool deliberately has no way to perform.
What happens to nested columns and structs?
They are flattened into dotted column names — an address struct becomes address.city and address.postcode. Repeated fields and arrays are written as JSON text in a single cell, because exploding them would change the row count.
Will my CSV open correctly in Excel?
Turn on "Add BOM for Excel" in the output options. Without it, Excel assumes the system encoding and mangles anything non-ASCII. If your data contains commas, Excel handles the quoting correctly either way.
How big can the Parquet file be?
The practical ceiling is your available RAM, not a server upload limit. Files in the low hundreds of megabytes convert comfortably on a normal laptop. Very large files are best split first, since the browser has to hold the parsed data in memory.
Keep going