Create

Summary

This page describes the Create page, where a new catalog is created: an index of the files of one folder, kept in the collection so that it can be searched and explored even when the device is disconnected.

It takes three steps:

1. Define the catalog: the folder to index, its storage device and its name.
2. Choose the content options: file types, sub-folders, hidden files, metadata, checksums and exclusions.
3. Click Create.

Catalog definition

Field	Description
Source path	The folder to index. It is filled with the path of the device selected on the Selection page, and can be typed or chosen with Select the path.
Storage	The storage device where the folder is: the drive, disk or computer (see Devices). Only storage devices are listed. A new collection already has one, Local disk. To add another one, use Add Storage on the Devices page.
Catalog name	The name of the new catalog. Each catalog needs its own name. Generate name from path (the button beside the field) makes one from the source path, with _ in place of / and :.

When a field is missing or wrong, a message at the top of the page says what is needed, for example Provide a name for this new catalog. It disappears when the field is corrected. The same happens when the source folder does not exist, or when another catalog already has the name.

When the source folder holds no file, a question is asked first: the folder may be empty, or the device may not be mounted there. Yes creates the catalog anyway, empty.

Content options

The options decide what is indexed. They can be changed later by editing the catalog (see Edit device); the catalog then needs an update to follow them.

File type

The catalog can be limited to one type of files.

Type	Description	Definition	Extensions examples
All	All file types, without filtering.
Audio	Music, podcasts, audio recordings and sound files.	As defined by MIME types.	MP3, FLAC, AAC, M4A, OGG, WAV, AIFF, Opus, WMA, MIDI, AMR (50+ formats)
Image	Photos, graphics, diagrams, icons and visual content.	As defined by MIME types.	JPG, PNG, HEIC, WebP, TIFF, RAW, SVG, XCF, GIF, BMP (100+ formats)
Text	Documents, code files, markup, data files and readable content.	Specific Katalog definition: all files with a MIME type starting with text/, plus application files like PDF and Word documents.	PDF, DOCX, ODT, Markdown, HTML, JSON, source code files, ebooks (100+ formats)
Video	Movies, clips, animations and video content.	As defined by MIME types.	MP4, MKV, AVI, WebM, MOV, FLV, 3GP, OGV, M2TS (40+ formats)
Other	All other file types.	Specific Katalog definition.	ZIP, RAR, EXE, DLL, ISO: executables, archives, system files and application files not classified as Text.
None	Files whose type cannot be found from the extension.	Files without extension or with an unknown extension.

The type of a file is found from its extension, using the MIME database of the system: hundreds of formats are recognised, and new ones as the system is updated. This keeps indexing fast. A file with a wrong or missing extension can therefore get the wrong type.

Include subdirectories

Ticked (the default), the files of all the sub-folders are indexed. Unticked, only the files directly in the source path.

Include hidden files

None (the default) skips hidden files and folders; All includes them.
On a home folder or a system drive, hidden folders (.cache, .local…) can hold many files: excluding the unneeded ones keeps the catalog small.

Include metadata

How much information is read from the content of the files. It makes indexing longer, but gives richer results in Search (dimensions, duration, artist…) and in Statistics.

Level	What is read
None	No metadata: fastest.
Media Basic	The essentials of images, videos and audio files: dimensions and orientation of images; dimensions, duration, codec and frame rate of videos; artist, album, track, duration and bit rate of audio files.
Media Extended	All the metadata of images, videos and audio files, including technical details.
Full Extended	All the metadata of every supported file type.

The extended levels use the KFileMetaData library, which decides the supported types and fields; their metadata can be seen with Show extended metadata (JSON) in the search results and in Explore.
Media levels read only media files: other files are skipped. Files that are damaged or cannot be read are skipped without stopping the catalog.

Supported media files:

• Images: jpg, png, gif, bmp, tiff, webp, svg, heic, raw, xcf
• Videos: mp4, mkv, avi, mov, wmv, flv, webm, m4v, mpg, 3gp, ogv, vob
• Audio: mp3, wav, flac, ogg, m4a, aac, wma, opus, aiff, mid, amr

Include checksum

SHA-256 calculates a checksum of each file: a fingerprint of its content. Checksums make the duplicate and difference searches exact, and allow verifying later that files did not change.
⚠️ It is much longer than the other options, because all the data of every file is read (see the Performance guide).

Exclude folders or files

Paths or names that are not indexed in this catalog.
Type a path or a name in Path or name to exclude, or choose a folder with Select the path, then click Add. The Remove button beside an entry takes it out of the list.
The list is saved with the catalog.

Global Parameters

The Global Parameters section holds settings shared by all the catalogs of the collection. The button on its title line folds or unfolds it.

Global exclude directory

Paths excluded from every catalog, when it is created and each time it is updated.
Type a path in Path to exclude globally (or use Select the path), then click Add. Remove beside an entry takes it out.

How exclusions work

Global and per-catalog exclusions add up: a file or folder skipped by either list is not indexed.
An exclusion is text matching: every file or folder whose full path contains the text is skipped. The matching is case-sensitive.

• Full path: /home/username/Downloads/temp skips that folder and all its content.
• Folder name: node_modules skips every node_modules folder, wherever it is.
• Part of a name: .cache skips /home/username/.cache/, but also /home/username/.cachedata/, since both paths contain the text.

Creating the catalog

Create saves the catalog and starts indexing. Progress is shown in the activity panel at the bottom of the window (see Interface). If another operation is running, the creation waits in the queue.

Indexing runs in two phases:

1. the list of files is read and saved: the catalog can be used as soon as this phase ends,
2. then, when chosen, the metadata and checksums are read.

Stop ends the creation:

• during the first phase, the catalog is not created,
• during the second phase, the catalog is kept with its full list of files; the metadata or checksums not read yet are completed by the next update of the catalog.

When the creation ends, the new catalog is selected on the Selection page and the Devices page opens with a report: the catalog name and path, its number of files and total size, and, when its storage was updated, the storage's used, free and total space with what was added. If a metadata or checksum scan was not finished, a message explains it instead.
Close (or Esc) closes the page.

Performance guide

What affects scanning speed?

#	Factor	Impact Level	What drives it	Notes
1	File Checksum	⚫⚫⚫⚫⚫ Critical	Total data volume (reads 100% of bytes, 8 MB buffer), then disk speed, then file count	SHA-256 via Qt QCryptographicHash. A 2 TB HDD at ~100 MB/s = ~5.5 hours just for I/O. SSD at ~500 MB/s ≈ 1 hour. Resumable if interrupted.
2	Storage hardware	⚫⚫⚫⚫ Very High	Sequential vs random I/O, throughput, latency	SSD ~100K files/min, HDD ~20–30K (fragmentation makes it worse). Network: highly variable. Affects ALL other factors.
3	Metadata: Full Extended	⚫⚫⚫ High	File count (especially video: 5–15 ms/file parsing containers), file type distribution	Uses KFileMetaData. Video/audio require seeking into file structure. ~10× slowdown vs None reported.
4	Metadata: Media Basic/Extended	⚫⚫⚫ High	File count of media files only	Reads headers only (images: ~2–3 ms). Non-media files skipped entirely — so file type mix matters a lot.
6	Include Hidden files	⚫⚫ Medium	Additional file count (can be massive on system/home drives: .cache, .local, node_modules…)	Multiplier effect: more files = more of everything above. Excluded directories are the mitigation.
7	Database mode	⚫ Low-Medium	Write throughput during indexing	Memory > File > Hosted. Memory avoids SQLite write I/O during scan; File mode is I/O-bound on writes. Batch transactions (1000 files/batch) already mitigate this.
8	File type filter	⚫ Low-Medium	Reduces file count processed for metadata	Filtering to Audio/Video/Image skips all Text/Other for metadata — significant if indexing a mixed drive.
9	File count (total)	⚫ Low	Drives filesystem walk time, DB insert count	Even without any options, enumerating 500K files takes time. Independent of sizes.
10	Excluded directories	⚫ Low	Reduces effective file count	node_modules, .cache, temp dirs can represent 20–30% of home drive file count.
11	System load / CPU	⚫ Low	Parallel metadata extraction uses 4–8 cores	Checksum is I/O-bound so CPU matters less there. Metadata is CPU-bound.

Other long processes, not used by Create or Update:

#	Factor	Impact Level	What drives it	Notes
5	MIME type verification	⚫⚫⚫⚫ Very High	File count (reads partial content of each file)	Update-only operation. "Guessing" from extension is free; verification reads file bytes. Impact scales with file count, not size.
12	Collection Export	⚫⚫⚫⚫ Very High	Export of tables to SQLite or csv

Test sample / data set: 60Gb, 40 000 files, with a file of 20Gb, and mix of text, images, video.

Test	Sample	SSD 2 TB (extrapolated)	HDD 2 TB (estimated)	Scaling basis
Baseline indexing	00:00:04	~00:02:37	~00:05–08:00	File count
Media Basic	00:01:23	~00:54:00	~02:00–03:00	File count (media files only)
Media Extended	00:01:52	~01:13:00	~03:00–04:00	File count (heavier per video)
Full Extended	00:01:23	~00:54:00	~02:00–03:00	File count
Checksum only	00:09:20	~06:06:00	~20:00–30:00	Total bytes / throughput
Checksum + Media Basic	00:11:06	~07:15:00	~22:00–33:00	Bytes (dominant) + count

Development

Some ideas of developments for this screen:

• Catalog a cloud instance (for example Nextcloud).
• Customize file types.
• For more, see the backlog of Create development.