Compare a stable copy of each list before deciding that anything is missing
When two lists look nearly identical, a missing product code or duplicated invitee can hide in a long export. Start by saving an untouched copy of each version, then paste comparable text into the Text Diff tool. Its line mode highlights added and removed lines, while word and character modes help inspect a small change within one line.
The comparison result is not a database reconciliation. It shows how the text you pasted differs and in what order. A reordered line can look like a deletion plus an addition, and different spacing, capitalization, or an extra header can make otherwise equivalent entries appear changed. Use it to create a review list, then confirm important findings in the source system.
Imagine comparing yesterday's and today's inventory export. SKU-1042 seems to be missing from today's list, while SKU-1043 appears twice. Before treating that as a stock issue, verify that both exports cover the same warehouse, date range, status filter, and column. A different export filter is a much more common explanation than an unexplained deletion.
Prepare the lists so the comparison answers one question at a time
Decide what one line represents before pasting anything. For an invite list, that may be one email address; for an inventory file, one SKU; for a report, one complete record. Remove titles, footers, timestamps, and notes that are not part of the comparison. If columns are involved, keep the same column order in both copies.
- Put the earlier or expected list on the left and the later or actual list on the right, and label the copies outside the tool so they cannot be swapped.
- Start with line mode to spot records that were added, removed, or moved. Use word mode for a changed quantity or suffix, and character mode only when a small spelling or punctuation change matters.
- Keep original line endings and avoid manually "fixing" values before the first pass; otherwise you can erase the evidence you need to review.
- If order is not meaningful, make separately sorted copies using the same rule, compare those, and keep the unsorted copies for investigating sequence changes.
For a CSV export, first check that each record really occupies one line and that a quoted field has not introduced unexpected line breaks. The CSV Preview tool can help you inspect column placement before you turn records into comparison lines. It does not validate your target system's import rules, so test a small representative file where that system allows it.
Separate reordered records from genuinely missing or repeated values
Suppose a supplier sends a revised list of 300 SKUs. A line diff might show twenty removals followed by the same twenty additions because that block was sorted differently. Search for one supposedly removed SKU in the right-hand list and one supposedly added SKU in the left-hand list. If both are present, the question is ordering, not presence.
For a potential duplicate, search for the exact identifier and then inspect its surrounding fields. Two rows with the same product name may be deliberate variants, while two rows with the same SKU and conflicting quantity need a source check. A text diff does not count unique values or understand a primary key; it only compares sequences of text.
It can help to make two review notes: "present in both but changed," and "present in only one copy after normalization." Include the row value, the source filename, the export time, and the filter used. This turns colored highlights into questions that another person can reproduce instead of a vague claim that the list is wrong.
Confirm the result before you update an invitation, order, or record
Never overwrite a master list based solely on a diff. Open the authoritative source, confirm the specific identifier, and check whether a recent status change, archived record, permissions change, or export setting explains it. If the list contains personal information, use redacted samples whenever possible and follow your organization's access rules.
For the inventory example, compare the expected SKU against the warehouse system, then check whether the duplicate line refers to a bundle, a retry, or an actual duplicated export row. Only after that confirmation should you correct the data or ask the owner of the source system to do so. Keep the two source copies and your comparison note until the issue is resolved.
As a final check, re-export a small affected slice and compare it again. Make sure the intended SKU is present once, the changed fields match the approved source, and unrelated lines did not move or disappear. The tool can make a review faster, but it cannot tell you which system is authoritative or whether a business change was approved.
Frequently asked questions
Does a removed line always mean the item was deleted?
No. It can mean the line was reordered, filtered out, reformatted, or pasted differently. Search for the identifier in both lists and confirm the export scope before concluding that it was deleted.
Can I use this to find duplicates automatically?
It can make repeated lines visible when you compare versions, but it does not calculate unique values or enforce a key. For important duplicate checks, use the source system or a controlled data workflow and verify the results.
Which mode should I use for two lists?
Start with line mode because each record is usually a line. Switch to word mode to inspect a changed field inside a record; use character mode for a very small text change such as an extra digit or punctuation mark.
What if the lists are the same entries in a different order?
Make copies sorted by the same stable identifier and compare those copies for presence. Preserve the original order too, because sequence changes may matter for a queue, priority list, or audit trail.
Is it safe to paste customer or production data into the tool?
Avoid unnecessary sensitive data. Use the smallest redacted sample that can reproduce the issue, remove credentials and personal details, and perform the final verification in the authorized source system.