Every CSV has too many columns, the wrong column names, or both. A CRM export from Salesforce might have 140 columns when your analysis needs 8. A supplier product feed has internal codes that should never be shared externally. A data handoff to a partner requires columns in a specific order with specific names. These are not complex data engineering problems — they're routine column management tasks that shouldn't require writing code.
Selecting: Keep Only the Columns You Need
The first and most common operation is simply choosing which columns to keep and discarding the rest. This makes files smaller, faster to work with, and easier to share. It also reduces the surface area for data quality errors — every column you keep is a column someone might misinterpret.
When selecting columns for a specific recipient, think about what they actually need. A finance partner needs amounts and dates, not product descriptions. A marketing team needs contact fields, not internal account IDs. Give each audience exactly what serves them.
Dropping: Removing Sensitive and Irrelevant Columns
Dropping is the inverse of selecting — you start with everything and remove what you don't want. This is the right approach when most columns are useful but a few should be excluded: PII columns before sharing externally, internal scoring fields before exporting to a partner, or deprecated columns that are always blank.
Sohovi automatically detects PII in your datasets — emails, phone numbers, SSNs — all processed client-side so your data never leaves the browser.
Pay special attention to columns that contain personal data. A standard CRM export often includes birth dates, phone numbers, national IDs, or salary information that should not be in a file sent to a third-party vendor or marketing platform. Dropping these columns before export is a data minimisation best practice under GDPR and CCPA.
Renaming: Standardising Column Headers
Column names from system exports are often system-internal identifiers: Account_Name__c, ContactID__r, custno, FIRST_NM. These mean nothing to the people receiving the data. Renaming maps technical field names to plain-English labels that match what the downstream system or person expects.
Consistent naming also prevents errors. If your file has a column called "email" and the import template expects "Email Address", the import will fail or map incorrectly. Rename before sharing to match the target system's expected headers exactly.
Reordering: Logical Column Sequence
Column order matters for readability. A file where name, email, and phone are in columns 1, 2, and 3 is much easier to review than one where they're in columns 1, 47, and 89. Reordering also matters for systems that process columns positionally rather than by name.
A browser-based CSV column picker handles all four operations in one interface: select which columns to keep, drag to reorder, click to rename, and download the result. No installation, no code, and your data stays in your browser.
Sohovi's free CSV Column Picker supports all four operations and processes your file locally.
Keep Reading
Try the Free CSV Column Picker →