Upload CSV, Excel and other files to Mammoth
Upload files from your computer to create datasets: open the Data Library, click New dataset, keep Upload selected, add your files and click Submit. You can upload one file, several files or a whole folder, as long as each upload is 50 MB or smaller. Mammoth detects the file type, structure and column types after you submit, and asks you when it can't decide on its own.
Supported file formats for upload
| Kind | Formats | How Mammoth reads them |
|---|---|---|
| Data files | Plain text (.txt), comma-separated (.csv), tab-separated (.tsv), pipe-separated (.psv), Microsoft Excel (.xls, .xlsx, .xlsm), XML (.xml), JSON (.json), JSON Lines (.jsonl, .ndjson) | Parsed straight into datasets with automatic column and type detection. A .txt delimiter is auto-detected; custom delimiters are supported. JSON Lines is one JSON object per line. |
| Compressed files | ZIP archive (.zip), Bzip2 compressed (.bz2), Gzip compressed (.gz), Tar archive (.tar), 7-Zip archive (.7z) | Extracted during upload; each file inside is read by its own format |
| Documents and images | PDF (.pdf), TIFF image (.tiff, .tif), JPEG image (.jpeg, .jpg), PNG image (.png), HEIC image (.heic, .heif), WebP image (.webp) | AI-assisted extraction turns their content into tables. See Extract tables from PDFs. |
- An Excel file with several sheets shows an Extract sheets action after upload.
- XML and JSON are read as a tree, and Mammoth picks the repeating records as rows. When more than one set of records could be the rows, it asks you to choose with Use as rows (see Review how Mammoth reads a file).
- An archive that holds one file becomes one dataset. One that holds several files becomes a folder named after the archive, with one dataset per file; if that folder name exists, Mammoth adds a number to the new folder's name.
To pick the best format, see Prepare files for a clean upload.
Upload a single file
- In the Data Library, click New dataset.
- Select Upload (the default option).
- Drag your file into the upload area, or click Upload file.
- Click Submit.
The dialog closes and the file appears in the Data Library as a pending item while Mammoth reads it, scans it for security risks and creates the dataset. A progress header shows the upload count ("Uploading 3 items", then "Upload complete"). You can cancel uploads that are still running.
Upload several files or a folder
Select several files, or click Upload folder, in the same dialog. Each file becomes its own dataset. See Upload several files or a whole folder for folder structure, folder rules and the 1,000-file limit.
What you see after the upload
Open the new dataset from the Data Library to see its panel. Overview holds a summary, actions (such as Explore and Add data), a Preview of your data, and Settings & history. That holds Settings (how the file was read, plus Update method, Unexpected columns and Cell editing for future batches), Batches (every load of data into this dataset; see Batches), Automations that use this dataset, and History of what has happened to it. See The dataset panel.
When Mammoth needs your input before it creates the dataset
If Mammoth can't finish on its own, the dataset's panel says "We need your input before we process this file" and asks for one of these. Each is explained in Review how Mammoth reads a file.
| Prompt | Why it appears |
|---|---|
| Choose how to read it | The file has more than one plausible way to be read |
| Extract sheets | An Excel file has several sheets ("This file has multiple sheets to be extracted.") |
| Which way round are these dates? | Mammoth can't tell whether dates are day first or month first |
| Extract Data | A PDF or image: choose what to import (see Extract tables from PDFs) |
| Access to this file is protected by a password | Enter the File password |
| This file seems to be corrupt | The file can't be turned into a dataset |
If the virus scan flags a file, Mammoth rejects it as a dangerous file and doesn't import it.
Limits
| Limit | Value |
|---|---|
| Upload size | 50 MB per upload request. Each file you select is its own request; a folder is one request. Compressing is a way to fit larger data. |
| Files per upload | Up to 1,000, sent in groups of 50 with a 30-second pause between groups |
See Limits and quotas.
Troubleshooting file uploads
| Symptom | Cause | Fix |
|---|---|---|
| "Upload failed", or the spinner never finishes | The upload is over 50 MB, the format isn't supported, the file is damaged, or the connection dropped | Check the total is under 50 MB and the format is in the list above. Download and re-upload the file. Check your internet connection. Contact support if a supported file under the limit still fails. |
| "File couldn’t be uploaded due to possible corruption or security risks" | The file failed the security scan | The file can't be imported. Get a clean copy from its source. |
| Data is in one column, characters or dates look wrong, or the wrong sheet came in | The file was read in a way you didn't expect | See Fix files that read wrongly |
Related
- Prepare files for a clean upload: pick a format and avoid read errors.
- Review how Mammoth reads a file: fix a messy read, pick sheets, answer date prompts.
- Add or replace data in a dataset: add a newer file to a dataset you already have.
- Explore cards: look at distributions and quality once the data is in.
Something wrong? support@mammoth.io