Diskover Search Cheat Sheet

Finding and tracking data — a quick reference. Applies to Diskover 2.6.0+.

Diskover indexes the metadata of every file — name, path, size, dates, owner, group, extension, and tags — so you can locate data across millions of files in under a second. You search the catalog, not the file contents. There are three ways in: Quick Search (one-click presets), Search Filters (a point-and-click builder), and the Search bar (typed queries). They all produce the same kind of result list.

The three ways to search

Search bar — the rules

1. Bare words search name, path, and extension at once (tumor finds it anywhere in the name or path). Internally this covers both the case-sensitive and case-insensitive name and path fields, so plain words are case-insensitive.

2. Multiple words are AND-ed automatically. Use OR / NOT and parentheses ( ) to change the logic (lowercase or/not/and are fine).

3. Case-insensitive by default — until you target a field. Field searches are case-sensitive except name.text and parent_path.text.

4. Wildcards are OFF by default. Turn them on with * (any characters) or ? (one character), e.g. sample*.csv.

5. Target a field with field:value, e.g. extension:fastq, size:>1073741824, owner:jsmith.

6. The Current Dir toggle (top right) limits a search to the folder you are in; turn it off to search the whole index.

Fields you'll use most

Operators, dates, and sizes

Operators: AND OR NOT combine terms · ( ) group logic · * many characters · ? one character · > < >= <= for size/count · [a TO b] inclusive range · \ escape a character or force a raw query.

Date math (mtime / atime / ctime): now (this instant) · now-7d (7 days ago) · now-3M (3 months) · now-1y (1 year) · [now-30d TO now] (last 30 days) · [* TO now-1y] (older than 1 year).

Size — you can type a raw byte count or a unit (Added 2.6.0: the search bar converts units to bytes for you):

Units accepted: b, kb, mb, gb, tb (case-insensitive). Raw byte counts still work everywhere.

Smart Search shortcuts

Press ! to start a type shortcut, / to jump to a path, or start with \ to force a raw Elasticsearch query.

!img images · !vid video · !aud audio · !doc documents · !code code · !datab databases · !compress archives · !exe executables · !sys system · !web web files · !tmp temp files · !discimg disc images

Added 2.6.0: this list is the shipped default and is editable — an admin can add or change shortcuts under Analytics ▸ Smart Searches (each is a name plus an Elasticsearch query). Your site's list may differ.

Research recipes — copy, paste, adapt

1. All my structured data / instrument exports — extension:(csv OR json OR dat OR xml) AND type:file

2. Locate a dataset by partial name — name:*patient042* AND type:file

3. Raw sequencing / imaging files by extension — extension:(fastq OR bam OR fasta OR tif OR czi) AND type:file

4. Everything inside one study folder (and subfolders) — parent_path:\/mnt\/research\/study_2024* (drop the * to stay in just that folder)

5. Large files — archive candidates — size:>1gb AND type:file (Quick ▸ Files ▸ Large files)

6. Stale data (untouched > 1 year) — mtime:[* TO now-1y] AND atime:[* TO now-1y] AND type:file

7. This week's work — mtime:[now-7d TO now] AND type:file

8. Empty / truncated files (failed runs) — size:0 AND type:file (Quick ▸ Empty files)

9. A colleague's files in a shared area — owner:jsmith AND parent_path:\/mnt\/shared*

10. Images only — !img (or spell it out: extension:(tif OR tiff OR png OR jpg))

11. Data tagged for QC — tags:"needs QC" AND type:file

12. The power search: recent, mid-size results — extension:(csv OR json) AND size:>1kb AND mtime:[now-90d TO now] AND type:file

13. Folders only (e.g., to size subtrees) (Added) — type:directory AND parent_path:\/mnt\/research\/study_2024*

14. Files with no extension (Added) — NOT extension:* AND type:file

15. Duplicate candidates by hard links (Added) — nlink:>1 AND type:file (for content duplicates, use Quick ▸ Duplicate files or the Dupes Finder)

Working with your results

Tag data for later

1. Run a search. 2. Tick the rows you want (or Select all). 3. Click Tag, pick or type a tag. 4. New-tag format: tag name|#hexcolor (e.g. needs QC|#e07b39). 5. Recall anytime with tags:"needs QC"; click a tag again to remove it.

On the results page: Edit/Save the query (save to Quick ▸ Saved Queries) · extension chips to narrow by type · search within results · Sort / Sort2 (two-level ordering by size, date, name) · Export the file list as CSV · left path tree to jump between top paths and drill into folders.

Gotchas

• You only see what you're allowed to see. Results are scoped to your Index Access rules and permission filters — empty results can mean "no access," not "no data." Check with an admin if a path you expect is missing.

• Escape slashes in typed path queries. Use \/ in parent_path: values; an unescaped / in a typed query can error out. (Building the query in Search Filters and reading what it wrote is the safe way.)

• Keyword fields are case-sensitive. name, extension, owner, tags match exactly (use name.text for case-insensitive name search); bare words are case-insensitive.

Pro tip

The full, always-current syntax reference lives on your Diskover server at Help ▸ help.php. Build a query in Search Filters, then read what it wrote into the search bar — the fastest way to learn the syntax.