Healthcare and HIPAA teams
Claims files and EHR extracts routinely run to millions of rows, arrive in inconsistent formats, and cannot be uploaded to an online tool. BigXL opens them at any size, cleans and merges them without a formula, and can apply HIPAA Safe Harbor de-identification before a file is shared. It makes no network requests for file processing, so protected health information stays on your machine.
The file is the problem before the analysis is
A year of claims lines does not fit in a spreadsheet. An EHR extract comes out with dates in three formats and member IDs already truncated by whatever opened it last. The obvious fixes all involve uploading the file somewhere, and that is exactly what you cannot do.
BigXL is built for that position: the file opens, the cleanup happens with menus rather than formulas, and nothing touches the network.
Open it
CSV, TSV, compressed .gz and .zst, Parquet, and Excel workbooks, at any
size, with no row limit. Files of 50 GB have been tested. Scrolling, sorting,
filtering, summary statistics and pivot tables are free.
Clean it
The recurring damage in health data is predictable, and the Pro cleaning tools handle it without a formula. Member and account numbers stay as text instead of collapsing into scientific notation or losing their final digits. Mixed date formats standardise to one pattern, and genuinely ambiguous dates are refused rather than guessed at. Columns that should be numeric but hold stray text are grouped so you decide what happens to each bad value.
Merge it
Combining months or sources by copy and paste shifts values into the wrong columns the moment a header is renamed or reordered, and nothing warns you. BigXL merges on the header name, keeps columns that appear in only some files, and previews the result before writing.
De-identifying before you share
The Enterprise tier includes a HIPAA preset built on the Safe Harbor method at 45 CFR 164.514(b)(2). It applies the 18 identifier categories in one pass, truncates dates to the year, masks ages of 90 and over to 90+, and replaces the 3-digit ZIP prefixes from the 17 low-population areas named in the regulation with 000**.
Masking is non-destructive. The source file is never modified, and masks can be undone from the timeline unless your organisation has made them irrevocable.
You can also save a privacy scan, which lists which columns appear to hold personal data and whether each is currently masked. The report contains no values from your file, only column names and counts, so it can be sent to a reviewer who is not cleared to see the underlying data.
What BigXL does not claim
The masking, hashing and detection features exist to help you meet your own obligations. They are not legal advice, they are not a certification, and BigXL does not warrant that detection finds every identifier in every file. Deciding whether a file is safe to share, and checking it, remains yours. The certificate BigXL can generate records the transformations applied with SHA-256 hashes of the source and export. It is BigXL’s own record of what happened, not an official certification, and it makes no compliance determination.
Common questions
Can I use an online CSV tool on a file containing PHI?
Doing so means uploading the PHI to a third party, which is usually the thing you are trying to avoid and may not be permitted under your agreements. BigXL processes files entirely on your own machine. Its optional update check sends only your signed licence key, never PHI.
What does the HIPAA preset actually do?
It targets the 18 identifier categories in the Safe Harbor method at 45 CFR 164.514(b)(2). Dates are truncated to the year, ages of 90 and over are masked to 90+, and 3-digit ZIP prefixes from the 17 low-population areas named in the regulation are replaced with 000**. Detection validates cell values rather than trusting column names.
Does that make my file HIPAA compliant?
No. The tool helps you apply the Safe Harbor rules, but compliance is a determination you make, not something software confers. BigXL does not warrant that automatic detection finds every identifier in every file, and you remain responsible for reviewing a file before it is shared.
Can masked data still be analysed?
Often, yes. Consistent hashing maps the same input to the same output every time, so masked data sets still join and counts still work, while the original values are irrecoverable. Date shifting moves dates by a consistent offset per record, so intervals within a record are preserved.