--- title: "Duplicates, and a scan that survives a real drive · Space Wizard" description: "A Duplicates tab next to the file browser, and a walk built for half a million directories and locked files." url: "https://royalsoftworks.com/products/space-wizard/features/duplicates-and-the-scan/" source: "https://royalsoftworks.com" format: "markdown" note: "Markdown rendering of the HTML page at `url`. Same content, same canonical URL." --- Features · 1 of 3 # Duplicates, and a scan that survives a real drive. A Duplicates tab next to the file browser, and a walk built for half a million directories and locked files. 3 Transports: SFTP, FTPS, FTP 3-pass Duplicate check: size → prefix → full hash 64 Directories read in parallel 100 GB Free-tier scan allowance Duplicates ### Spot duplicates while you browse. A Duplicates tab sits right next to the file browser. Space Wizard groups files by exact size, narrows the survivors with a 4 KB prefix hash, then confirms with a full SHA-1 — a 4 GB video that differs in its first kilobytes costs 4 KB of reading, not 4 GB. - Dedicated Duplicates tab alongside Files and Usage - Three passes: exact size → 4 KB prefix hash → full hash on what's left - Biggest wins first — groups arrive sorted by file size - Selecting every copy in a group is blocked; one always survives - Local deletes go to the recycle bin or trash, not straight out For your IT team - **Engine** — Size grouping → SHA-1 over the first 4,096 bytes → full SHA-1, 8 files hashed concurrently - **Remote** — Identical pipeline; the hashes are computed on the remote host instead of over the wire - **Deletion** — shell.trashItem() locally; remote deletes are permanent and labelled as such in the UI The scan itself ### Built to survive a real system drive. Half a million directories, junctions pointing at each other, files the kernel won't let you stat — a walk that gives up on any of those reports a number that's quietly wrong. This one doesn't. - 64 directories read at once, results streaming in batches as they land - pagefile.sys and other locked files are sized from the directory entry — roughly 15 GB that most scanners simply lose - Junctions and symlinks resolved so nothing is counted twice, and links back into the tree are skipped - Optional: stay on one filesystem, follow links, include cloud folders - Cancel at any point — a stopped scan keeps whatever it already measured For your IT team - **Walk** — Concurrent breadth-first, 64-way, with per-directory 'direct size' totals summed in the renderer - **Locked files** — Windows: Get-ChildItem reads the size from the directory entry when open() is refused (EBUSY/EPERM) - **Elevation** — Asks for Administrator rights at launch; any directory that still can't be read is counted and reported [Previous Features: all 3 pages](https://royalsoftworks.com/products/space-wizard/features/) [Next One window, three tabs.](https://royalsoftworks.com/products/space-wizard/features/one-window/) [All of Space Wizard features](https://royalsoftworks.com/products/space-wizard/features/)