Guides

Know what a table change means.

Short, practical notes on reading a file, choosing an explicit rule, and reviewing what the export will change.

Start here

A five-minute check before your next CSV import

Read the structure first. A clean-looking spreadsheet can still carry an extra separator, an empty header, or a blank required value.

Cleaning

Clean text without damaging identifiers

Whitespace is often accidental. Leading zeroes, case, and internal spacing can be meaningful. Apply each rule to the right columns.

Duplicates

Choose a duplicate key you can defend

Two records with the same name are not automatically the same customer. A useful duplicate rule starts with the identity of the record.

Comparison

Compare two snapshots without false matches

A dependable comparison needs a stable key, a clear before/after order, and an explicit meaning for a change.

Dates

Make date formats explicit before converting

04/03/2026 can describe two different days. A useful formatter asks you to choose the source convention rather than guessing.

Validation

Build an import checklist from actual rules

A structurally valid CSV can still violate the receiving system’s rules. Turn those requirements into explicit checks.

Filtering

Filter records with one explainable rule

A filtered export is easier to trust when the field, operator, and comparison value are all visible.

Sorting

Sort a table without converting its values

The sort type determines the order. The exported strings remain the same as the source.

Column selection

Choose columns before sharing a table

A smaller export can be easier to review, but a smaller file is not automatically a safe file.

Replacement

Make find-and-replace changes reviewable

A literal search is safer to reason about than a broad text rule, but every replacement still deserves a before-and-after check.

Splitting

Split a field using a literal separator

A delimiter inside a parsed cell is ordinary text. Choose the exact separator your source uses.

Combining

Combine fields with a visible separator

Joining fields is convenient for display, but it can remove boundaries that were clear in the original columns.

Filling

Fill missing cells without hiding the rule

A fill operation is a data decision. Write down the rule that makes the replacement valid for this column.

Mapping

Map categories with an explicit dictionary

A small lookup list makes category changes reproducible and visible.

Frequency

Read exact-value frequency tables

Counts help reveal repeated labels and unexpected categories. Exact counts are only as meaningful as the value normalization you choose.

Sampling

Make a repeatable random row sample

A fixed seed lets a reviewer reproduce a subset from the same ordered file.

File splitting

Split a CSV into ordered parts

Record-based splitting keeps quoted multiline fields together and makes each part independently readable.

Appending

Append tables by header name

Appending is a vertical stack. Header matching prevents a different column order from moving values into the wrong fields.

Joining

Join two tables with unique keys

A safe join begins with keys that identify one row on each side.

Statistics

Interpret column statistics in context

A profile is a quick description of observed values, not a verdict on data quality.

Grouping

Summarize numeric values by category

A group summary converts many rows into one row per exact category.

Formats

Choose a table export format

Choose an output format for the next application while keeping the limits of flat tables in view.

Pivoting

Build a two-way summary from CSV categories

A two-way summary answers how one category breaks down across another without changing the source records.

Reshaping

Reshape a wide CSV table into long records

A wide table stores repeated measures in separate columns. A long table stores the measure name and value in rows.

Column order

Reorder CSV columns for a specific import

Some receiving systems expect a known field order. Rearrange headers and their complete columns without editing their values.

Examples use fictional data. Check each workflow against the requirements of your own destination system.