URLNotesGrabberCore/URLNotesGrabberCORE/TL.db.md

TL.db — schema notes

The SQLite database behind URLNotesGrabberCORE and its sibling crawlers, and the one Rolodex reads.

Everything below was read out of the live file, not inferred from code. Counts are as of 2026-08-07; re-run the queries at the bottom to refresh them.

• Journal mode: WAL — TL.db-wal and TL.db-shm live beside the file and are part of the database. Copying TL.db alone gives you whatever was last checkpointed, not the current state.
• Page size: 4096. File size: 148 MB.

⚠ Breaking change, 2026-08-07: Notes holds integer IDs, not names

Notes.RootBlogName, Notes.NoteBlogName and Notes.Type no longer exist. They are now RootBlogId, NoteBlogId and TypeId, resolved through the new BlogNames and NoteTypes tables. Any query naming the old columns fails outright.

There is no compatibility view. See porting to the integer schema for the old-to-new translation of every query shape the applications use.

Applied by ../normalize-notes.sql, which took the file from 207 MB to 148 MB. An earlier change the same day (../shrink-db.sql) took it from 267 MB to 207 MB.

---

The three content tables

Table	Rows	What it is
Blogs	188,620	The crawl registry — one row per known blog, plus crawl-state flags
Posts	22,468	Stored post content. Only 3,867 blogs actually have any
Notes	1,182,333	The engagement graph: NoteBlogId acted on (RootBlogId, PostID)

…supported by two lookup tables that exist only to keep Notes small:

Table	Rows	What it is
BlogNames	20,430	BlogId ⇄ BlogName. The ID authority for everything in Notes
NoteTypes	5	TypeId ⇄ Type. like, reblog, reply, posted, post_attribution

The engagement graph is the interesting part. 20,311 distinct blogs appear as engagers — far more than the 3,867 that have stored posts — which is what makes this a social graph rather than a post archive. Only 2,771 blogs appear as the root of a note.

Blogs

CREATE TABLE "Blogs" (
    "BlogName"              TEXT,
    "HasBeenOutput"         INTEGER DEFAULT 0,
    "IsActive"              INTEGER DEFAULT 1,
    "DateAdded"             TEXT NOT NULL DEFAULT '12/24/25',
    "ByLikes"               INTEGER NOT NULL DEFAULT 0,
    "LikesPulled"           INTEGER NOT NULL DEFAULT 0,
    "LikesCursor"           INTEGER DEFAULT 0,
    "DateModified"          TEXT,
    "DateCreated"           TEXT,
    LikesNewestTimestamp    INTEGER DEFAULT 0,
    LikesLastRefreshed      INTEGER DEFAULT 0,
    LikesLastNewCount       INTEGER DEFAULT 0,
    TTFolderPath            TEXT,
    BlogId                  INTEGER,
    PRIMARY KEY("BlogName")
);

CREATE INDEX ix_Blogs_BlogId ON Blogs (BlogId);

BlogName is the primary key, so it is the only indexed way in by name. There is no index on any flag or date — filtering or sorting on those scans all 188k rows, which is affordable here and is not on Notes.

BlogId is new as of 2026-08-07 and is the join key to Notes. It exists so that Notes can reach Blogs in a single integer hop rather than going through BlogNames and ending in a text comparison:

-- what you want
FROM Blogs B JOIN Notes N ON N.NoteBlogId = B.BlogId

-- not this
FROM Blogs B JOIN BlogNames BN ON BN.BlogName = B.BlogName
             JOIN Notes N      ON N.NoteBlogId = BN.BlogId

BlogId is NULL on 168,202 of 188,620 rows — every blog that has never appeared in a note. That is the large majority, and it is not an error: the registry is far bigger than the engagement graph. An inner join on BlogId therefore silently drops those blogs, which is usually what you want for engagement queries and is wrong for registry listings.

Flag distribution: IsActive = 1 on 188,601 of 188,620 rows, HasBeenOutput = 1 on 5,059, ByLikes = 1 on 2. IsActive carries a second meaning as of Rolodex — see Blogs.IsActive below.

The columns after DateCreated were added later by ALTER TABLE, which is why they carry no quoting in the stored DDL. That is the normal way this schema grows, and BlogId is the newest example.

DateAdded is not written consistently. 170,677 rows hold ISO yyyy-MM-dd HH:mm:ss; 17,943 hold US-format M/d/yy from a bulk import. As text those two sort into different parts of the table, so anything ordering or range-filtering on this column has to normalise first — see DateSql in Rolodex.

Posts

CREATE TABLE "Posts" (
    "BlogName"              TEXT,
    "PostID"                INTEGER,
    "HasNotesGathered"      INTEGER DEFAULT 0,
    "reblogURL"             TEXT,
    "NotFound"              INTEGER DEFAULT 0,
    "PostDate"              TEXT,
    "NotesGatheredDateTime" INTEGER NOT NULL DEFAULT 1729746000,
    "HasImage"              INTEGER NOT NULL DEFAULT 0,
    "PostURL"               TEXT,
    "Slug"                  TEXT,
    "ReblogKey"             TEXT,
    "ReblogName"            TEXT,
    "Summary"               TEXT,
    "Quote"                 TEXT,
    "Body"                  TEXT,
    "Tags"                  TEXT,
    "Link"                  TEXT,
    "PhotoURL"              TEXT,
    "PhotoCaption"          TEXT,
    "DownloadedFiles"       TEXT,
    "AudioCaption"          TEXT,
    "Question"              TEXT,
    "Answer"                TEXT,
    "Title"                 TEXT,
    "ByLikes"               INTEGER NOT NULL DEFAULT 0,
    "RootBlogName"          TEXT,
    "RootURL"               TEXT,
    "DateModified"          TEXT,
    "DateCreated"           TEXT,
    PostType                TEXT,
    PRIMARY KEY("BlogName","PostID")
);

The key is (BlogName, PostID), not PostID. This matters more than it looks: 345 post IDs exist under more than one blog, so an ID on its own is both ambiguous and unindexed. Any lookup should carry the blog name, and a batch lookup should group by blog so it stays on the leading column of the key.

Notable:

• PostType is now mostly populated: 20,679 of 22,468 rows, leaving 1,789 NULL. This reverses what earlier revisions of this document said — the column really was empty on every row, and something has since started writing it. Anything that treated it as permanently unset, or derived the type from post content instead, should be re-examined against the live data. Rolodex still derives it.
• HasImage = 1 on 14,026 rows. It records that the post had a picture, not that a usable URL was kept, so it is not a reliable predictor that anything will render.
• PhotoURL is largely unused; in practice the image markup lives inside Body.
• NotFound = 1 on 4,712 rows — posts that have since been deleted upstream.
• The content columns (Body, Quote, Question, Answer, …) are the heavy ones. List views should not select them.

Notes

CREATE TABLE Notes (
    RootBlogId      INTEGER NOT NULL,
    PostID          INTEGER NOT NULL,
    NoteBlogId      INTEGER NOT NULL,
    TimeStamp       INTEGER NOT NULL,
    TypeId          INTEGER NOT NULL,
    replyText       TEXT,
    DatetimeCrawled TEXT,
    DateModified    TEXT,
    DateCreated     TEXT,
    IsActive        INTEGER NOT NULL DEFAULT 1,
    PRIMARY KEY (RootBlogId, PostID, TimeStamp, TypeId, NoteBlogId)
) WITHOUT ROWID;

CREATE INDEX ix_Notes_NoteBlogId ON Notes (NoteBlogId);

Integer IDs since 2026-08-07 — this is the breaking change. RootBlogName, NoteBlogName and Type are gone, replaced by RootBlogId, NoteBlogId and TypeId. Resolve them through BlogNames and NoteTypes, or join straight to Blogs on BlogId. The old names were text repeated on 1.18 million rows, in the table and in every index over it; the swap took the file from 207 MB to 148 MB.

The primary key column order is deliberately unchanged, so the leading-prefix access patterns callers already depend on still hold: (RootBlogId) and (RootBlogId, PostID) remain cheap prefixes, exactly as (RootBlogName) and (RootBlogName, PostID) were.

Two nulls-and-defaults differences from the old DDL, both intentional:

• replyText and DatetimeCrawled no longer carry column defaults. The old table defaulted them to '.' and '2/12/26 12am', which is how 1.1M rows acquired placeholder values nobody wrote. New rows now get NULL unless a writer supplies something. The crawler names both columns explicitly, so its behaviour is unchanged.
• The five key columns are now NOT NULL. They always were in practice.

WITHOUT ROWID, since earlier the same day. The rows live in the primary key's b-tree rather than in a rowid table with a separate key index beside it. Two consequences matter before adding an index here:

• There is no rowid on this table. SELECT rowid FROM Notes is an error, and no code in any of the three apps relied on it.
• A secondary index carries the whole five-column primary key as its row reference instead of a compact rowid, so indexes here are expensive — though far less so than before, now that the key is five integers rather than three integers and two strings. ix_Notes_NoteBlogId costs 27 MB; its text predecessor cost 58 MB.

Notes_idx_06e01ae3 on TimeStamp DESC was dropped at the same time. It cost 14 MB as a rowid index and would have cost 58 MB after the conversion. It was worth neither: the crawler's only TimeStamp filter (>= 1535778000) excludes 786 rows of 1.18M, Rolodex's default Notes sort carries a three-column tiebreaker that forces a full sort regardless, and the reply-matching UPDATE uses ABS(TimeStamp - ?) <= 5, which no index on TimeStamp can serve. The one path that got slower is Rolodex's Notes page with a date-range filter: 60 ms to 164 ms.

See ../shrink-db.sql for the full rationale and the applied result.

DatetimeCrawled is NULL on 1,148,077 rows, and that is the honest value. Those rows previously stored the literal string '2/12/26 12am' — this column's own DDL default, written as a bulk backfill placeholder rather than as a crawl time. They were set to NULL on 2026-08-07, which is what consumers already displayed them as: the string parses as a date in neither format this schema writes.

Note the trap: the DEFAULT '2/12/26 12am' clause is still in the DDL above. Any INSERT that omits this column writes the placeholder straight back. The crawler names it explicitly on every insert, so nothing reintroduces it today, but a new writer that forgets to would — which is why consumers should keep treating an unparseable value here as "unknown" rather than assuming NULL is now the only such marker.

One row per engagement event. TimeStamp is unix seconds — unlike every date column elsewhere in the schema, which are text.

At 1.18M rows this is the table that dictates how the whole database has to be queried:

• Nothing should run an unbounded SELECT or a bare COUNT(*) here. A count scans the lot on every call.
• The only fast access paths are the primary key's leading columns (RootBlogId, then PostID) and ix_Notes_NoteBlogId on NoteBlogId. "Notes received by a blog" and "notes given by a blog" are both cheap; almost nothing else is.
• Every ordering here is a full sort of whatever the filters leave, TimeStamp included. Filter first, then sort.
• replyText is '.' on 1,167,464 rows — only reply notes carry real text. Those dots are inherited from the old column default; new rows get NULL instead.

Resolve IDs by filtering the lookup, not by scanning Notes. The lookup tables are tiny and uniquely indexed, so pushing a name predicate into them costs nothing and lets the Notes index do the work:

-- good: BlogNames resolves the name, then the index is searched
SELECT * FROM Notes
 WHERE NoteBlogId = (SELECT BlogId FROM BlogNames WHERE BlogName = ?);

-- also good, same plan
SELECT n.* FROM Notes n
  JOIN BlogNames b ON b.BlogId = n.NoteBlogId
 WHERE b.BlogName = ?;

BlogNames

CREATE TABLE BlogNames (
    BlogId   INTEGER PRIMARY KEY,
    BlogName TEXT NOT NULL UNIQUE
);

20,430 rows — every name appearing in Notes as either participant, and nothing else. This is the ID authority: Notes.RootBlogId and Notes.NoteBlogId both point here, and Blogs.BlogId is a copy of the value for the blogs that have one.

12 of these names have no Blogs row. The registry has never been a superset of the engagement graph and still is not, so resolving an ID through Blogs rather than BlogNames will occasionally find nothing. Use BlogNames when you need the name itself and Blogs when you need registry columns.

IDs are assigned by SQLite and are stable: they are stored in over a million Notes rows. Never renumber them. A blog that is renamed upstream should get a new row, not an edit to an existing one, unless every Notes reference is migrated with it.

NoteTypes

CREATE TABLE NoteTypes (
    TypeId INTEGER PRIMARY KEY,
    Type   TEXT NOT NULL UNIQUE
);

TypeId	Type	Rows	Share
1	like	945,167	79.9%
2	reblog	219,203	18.5%
3	reply	15,345	1.3%
4	posted	2,617	0.2%
5	post_attribution	1	—

The set is fixed in practice, but it is a table rather than a CHECK constraint so that adding a type is an INSERT and not a schema migration. The IDs above are stored in Notes and must not be reassigned.

Five rows means the lookup is effectively free; write t.Type = 'reblog' and let SQLite resolve it, or hardcode the ID if you prefer — both are fine, but hardcoding ties your code to this table's contents, so prefer the join in anything long-lived.

---

Porting to the integer schema

Everything here was checked against the live 148 MB file. There were 14 affected call sites in DataAccess.cs and 16 in RolodexRepository.cs. TumblThree needs no changes — its single statement touches Blogs.IsActive and BlogName only.

DataAccess.cs is ported. All 14 sites now read the integer schema, AddNote registers names and types before inserting, and verify-db-schema.sql reports a pre-migration file rather than letting the app fail on it. RolodexRepository.cs lives in the Rolodex repository and is not covered by that work. One site was dropped rather than translated: the LEFT JOIN Notes in GetPosts selected nothing and was collapsed by the query's own GROUP BY, so it could not affect the result.

Column mapping

Was	Is now	Resolve via
Notes.RootBlogName	Notes.RootBlogId	BlogNames.BlogId → .BlogName
Notes.NoteBlogName	Notes.NoteBlogId	BlogNames.BlogId → .BlogName
Notes.Type	Notes.TypeId	NoteTypes.TypeId → .Type
ix_NoteBlogName01	ix_Notes_NoteBlogId	—

PostID, TimeStamp, replyText, DatetimeCrawled, DateModified, DateCreated and IsActive are unchanged.

Filtering by a blog name

-- was
WHERE NoteBlogName = @Name

-- now, either form; both search ix_Notes_NoteBlogId after a unique-index lookup
WHERE NoteBlogId = (SELECT BlogId FROM BlogNames WHERE BlogName = @Name)
-- or
JOIN BlogNames b ON b.BlogId = n.NoteBlogId WHERE b.BlogName = @Name

Measured 73 ms against 63 ms for the old text form on the busiest blog — the extra hop is a unique-index probe on a 20k-row table and does not show.

Joining Notes to Blogs

This is the join to get right; it is the most common shape in both applications.

-- was
FROM Blogs B INNER JOIN Notes N ON N.NoteBlogName = B.BlogName

-- now: one integer hop, using the new Blogs.BlogId
FROM Blogs B INNER JOIN Notes N ON N.NoteBlogId = B.BlogId

Do not route this through BlogNames — that adds a hop and ends in the text comparison the change was meant to remove.

Selecting a name back out

-- was
SELECT NoteBlogName AS blogName, COUNT(*) FROM Notes ... GROUP BY NoteBlogName

-- now
SELECT bn.BlogName AS blogName, COUNT(*)
  FROM Notes n JOIN BlogNames bn ON bn.BlogId = n.NoteBlogId
 ... GROUP BY bn.BlogName

Group by n.NoteBlogId instead of bn.BlogName when you only need the name for display — grouping on the integer is cheaper and the name comes along for free.

Filtering by type

-- was
WHERE type IN ('reblog', 'reply', 'posted')

-- now
WHERE TypeId IN (SELECT TypeId FROM NoteTypes WHERE Type IN ('reblog','reply','posted'))
-- or, equivalently
JOIN NoteTypes t ON t.TypeId = n.TypeId WHERE t.Type IN ('reblog','reply','posted')

WHERE TypeId IN (2,3,4) also works and is marginally faster, but hardcodes this table's contents into application code. Prefer the lookup outside of hot paths.

Note the negated form needs care: type NOT IN ('reblog','reply','posted') becomes TypeId NOT IN (SELECT TypeId FROM NoteTypes WHERE Type IN (...)), which is correct only because TypeId is NOT NULL.

Inserting a note

The crawler must ensure both names have IDs first. INSERT OR IGNORE on BlogNames is the whole of it — no read-back, no round trip, safe to run every time:

INSERT OR IGNORE INTO BlogNames (BlogName) VALUES (@rootBlogName);
INSERT OR IGNORE INTO BlogNames (BlogName) VALUES (@noteBlogName);

INSERT OR IGNORE INTO Notes
       (RootBlogId, PostID, NoteBlogId, TimeStamp, TypeId,
        DatetimeCrawled, DateModified, DateCreated)
SELECT (SELECT BlogId FROM BlogNames WHERE BlogName = @rootBlogName),
       @PostID,
       (SELECT BlogId FROM BlogNames WHERE BlogName = @noteBlogName),
       @TimeStamp,
       (SELECT TypeId FROM NoteTypes WHERE Type = @Type),
       @DatetimeCrawled, @DateModified, @DateCreated;

Verified: a genuinely new note inserts, and re-running the identical statement inserts 0. Run all three statements in one transaction so a crash cannot leave a name registered with no note.

The duplicate-key error message has changed. DataAccess.cs compares against the literal string

UNIQUE constraint failed: Notes.RootBlogName, Notes.PostID, Notes.TimeStamp, Notes.Type, Notes.NoteBlogName

at two call sites to decide whether to swallow an exception. SQLite now emits the new column names, so those comparisons no longer match and real errors will surface where they used to be silently ignored — or vice versa.

Both sites now go through IsNotesDuplicateKey in DataAccess.cs, which matches on UNIQUE constraint failed plus Notes. rather than on the column list. A literal comparison is what broke here; the next rename should not break it again.

Updating notes

Predicates translate the same way. The reply-matching update, which cannot use an index on TimeStamp either before or after:

-- now
UPDATE Notes SET replyText = @replyText, DateModified = @dateModified
 WHERE NoteBlogId = (SELECT BlogId FROM BlogNames WHERE BlogName = @noteBlogName)
   AND ABS(TimeStamp - @TimeStamp) <= 5
   AND TypeId = (SELECT TypeId FROM NoteTypes WHERE Type = 'reply')
   AND (replyText IS NULL OR replyText = '' OR replyText = '.')
   AND (replyText IS NULL OR replyText <> @replyText);

Rolodex's soft-delete updates need no change beyond the WHERE clause — they set IsActive, which is untouched.

Three traps

Blogs.BlogId is NULL on 168,202 of 188,620 rows. Any inner join on it silently drops every blog that has never appeared in a note. Correct for engagement queries; wrong for registry listings, which need a LEFT JOIN or no join at all.

12 names in BlogNames have no Blogs row. Resolving an ID to a name through Blogs will occasionally find nothing. Use BlogNames for names and Blogs for registry columns.

IDs are stable and must stay so. BlogNames.BlogId and NoteTypes.TypeId are stored in over a million Notes rows. Never renumber. A blog renamed upstream gets a new row, not an edited one, unless every Notes reference migrates with it.

---

Referential integrity

There are no foreign keys, and the tables do not perfectly agree:

• 4 Posts rows name a blog with no Blogs row.
• 12 of the 20,430 names in BlogNames have no Blogs row.

So a name appearing in Notes or Posts is not a guarantee that the registry knows about it. Joins from those tables back to Blogs should tolerate a miss.

The integer schema does not fix this and was not meant to. BlogNames is deliberately built from Notes rather than from Blogs, precisely so that the 12 unregistered engagers keep their IDs and their rows. Had it been built from the registry, those notes would have been dropped by the migration's inner joins.

---

The '.' placeholder convention

The crawler writes a single dot into text columns it has no value for, rather than NULL. This is the single most surprising thing about the schema and it affects every consumer.

Column	'.' rows
Notes.replyText	1,167,464
Posts.Title	12,562
Posts.Body	172

Notes.replyText and Notes.DatetimeCrawled no longer carry column defaults as of the integer migration, so new note rows get NULL rather than a placeholder. The dots already in replyText were not rewritten — cleaning is still required on read.

Any query whose output reaches a human should collapse it:

NULLIF(NULLIF(SomeColumn, '.'), '') AS SomeColumn

Empty string turns up too, hence the double NULLIF. Not every column is affected — Blogs.TTFolderPath and Blogs.DateModified currently have zero dot rows — but new columns tend to acquire them, so treat cleaning as the default for any text column rendered to a user.

---

Supporting tables

Crawler bookkeeping. Rolodex ignores all of these.

Table	Rows	What it is
DailyAPICount	133	(Date TEXT PK, APICount INTEGER) — per-day API call tally against the rate limit
ApiKeyPoolState	2	(KeyName TEXT PK, RetryUntil INTEGER) — per-key backoff; RetryUntil is unix seconds
ApiKeyPoolMeta	1	(Id PK CHECK (Id = 1), LastIndex) — round-robin cursor. Singleton by check constraint
CollectRunState	1	(Id PK CHECK (Id = 1), RunCutoff, RunComplete, RunStarted, RunCompletedAt) — resume state for an interrupted collection run. Also a singleton

---

Blogs.IsActive — now written by two applications

IsActive has always been the crawler's work-selection flag. GetBlogs in DataAccess.cs joins on it to decide what to collect:

-- shape only; the ported GetBlogs joins Blogs directly on BlogId and needs no BlogNames hop
SELECT bn.BlogName, count(*)
  FROM Notes n
  JOIN Blogs     b  ON b.BlogId  = n.NoteBlogId
  JOIN BlogNames bn ON bn.BlogId = n.NoteBlogId
 WHERE b.IsActive = @isActive AND ...

Nothing inside the crawler writes it — it is an input, set from outside.

Rolodex is now one of the things that sets it. Removing a blog through the Rolodex UI runs exactly this:

UPDATE Blogs SET IsActive = 0 WHERE BlogName = ?;

Rolodex adds no column and changes no schema. It reuses this flag because the two meanings were judged to be one decision: a blog you do not want in the browsing UI is a blog you do not want to keep crawling. Removal therefore stops collection, and the Rolodex confirmation screen says so before anyone commits.

• 1 (or absent/NULL) — live. Crawled, and visible in Rolodex.
• 0 — removed. Not crawled, hidden from the Rolodex registry, dashboard counts and engagement rollups.

Restoring is the same UPDATE with a 1. Nothing is destroyed either way: the blog's Posts and Notes rows are never touched, and Rolodex deliberately keeps showing them under its Posts and Notes pages. Removing a blog hides the blog, not what it collected.

What other tools need to know

1. Setting IsActive = 0 now also hides the blog from Rolodex, and setting it back to 1 makes it reappear. If another tool deactivates blogs in bulk, it is also removing them from the browsing UI — which may be exactly right, but it is no longer a crawler-only decision.
2. Re-crawling a removed blog will not bring it back, since nothing in the crawler writes the flag. An INSERT OR REPLACE on the Blogs row would, by resetting it to the column default of 1. Prefer an UPDATE of the specific columns, or INSERT … ON CONFLICT DO UPDATE SET naming only the columns being refreshed.
3. NULL is treated as live. The column is INTEGER DEFAULT 1 with no NOT NULL, so Rolodex reads it through COALESCE(IsActive, 1). A NULL therefore leaves the blog visible rather than stranding it outside both the registry and the removed list, where no screen could reach it. Write 0 or 1, not NULL.
4. Backing the feature out is a configuration change, not a migration. Because there is no Rolodex-owned column, setting Rolodex__EnableBlogDeletion=false is the whole of it; there is nothing to drop. Any blogs already at IsActive = 0 simply go back to being ordinary inactive blogs.

---

Posts.IsActive and Notes.IsActive — present, and written from outside

The same flag extends to the two content tables, with the same meaning: 0 is removed, anything else — including NULL — is live. Both columns now exist in the live TL.db and are included in the DDL quoted above. As of 2026-08-07, Posts.IsActive = 0 on 5,900 rows and Notes.IsActive = 0 on none. Like Blogs.IsActive, they are written from outside this crawler.

On Notes the column is INTEGER NOT NULL DEFAULT 1, so a NULL cannot occur there; Posts and Blogs are laxer, which is why the predicate below still uses COALESCE.

The crawler therefore treats both as optional, and as nothing it owns:

• It never writes them. No INSERT column list names IsActive, no UPDATE sets it, and MapPrefixToColumn — the only place a column name is chosen at runtime — cannot map to it. Re-crawling a removed post or note refreshes its content and leaves the flag at 0. There is no INSERT OR REPLACE on Posts or Notes for a default to be reset by.
• It filters on them only when they exist. HasIsActiveColumn in DataAccess.cs asks PRAGMA table_info once per table per database path and caches the answer; the filter is COALESCE(IsActive, 1) = 1, and it is dropped entirely when the column is absent. Naming a missing column is a hard SQLite error, so this is what lets one build run against databases on both sides of the change. The cache lives for the process — adding the columns to a live database takes effect on the next run.

Every read that selects posts or notes carries the filter: GetPosts, GetReplies, GetRepliesWithMissingText, GetRepliesWithFilledText, GetAllPostTextColumns, GetAllPostsForBlog, GetPost, GetPostByIdAnyBlog, and the engagement queries that count or join Notes (GetBlogs, GetBlogsAll, GetBlogsForLikes). The one deliberate omission is the LEFT JOIN Notes in GetPosts: nothing is selected from it and it can neither add nor remove a row, so filtering it would buy nothing.

LegacyPostsDbImporter is unfiltered too — it reads a foreign legacy database whose Posts table is not this schema.

Two consequences worth stating plainly, both inherited from how Blogs.IsActive is handled:

1. Removal hides a row; it does not freeze it. The write paths are keyed on a post the caller already selected, so an ingest or a correction run still overwrites the content of a removed post. Only selection is filtered.
2. NULL is live. Write 0 or 1, not NULL, but a NULL leaves the row visible rather than stranding it.

---

Reproducing the numbers

SELECT 'Blogs', COUNT(*) FROM Blogs
UNION ALL SELECT 'Posts',     COUNT(*) FROM Posts
UNION ALL SELECT 'Notes',     COUNT(*) FROM Notes
UNION ALL SELECT 'BlogNames', COUNT(*) FROM BlogNames;

-- note type mix (joins NoteTypes; Notes.Type no longer exists)
SELECT t.Type, COUNT(*)
  FROM Notes n JOIN NoteTypes t ON t.TypeId = n.TypeId
 GROUP BY t.Type ORDER BY 2 DESC;

-- how much of the registry participates in the engagement graph
SELECT COUNT(*) FILTER (WHERE BlogId IS NOT NULL) AS with_notes,
       COUNT(*) FILTER (WHERE BlogId IS NULL)     AS without_notes
  FROM Blogs;

-- the two date shapes in Blogs.DateAdded
SELECT CASE WHEN DateAdded LIKE '____-__-__%' THEN 'ISO' ELSE 'US' END, COUNT(*)
FROM Blogs GROUP BY 1;

-- post IDs that are ambiguous without a blog name
SELECT COUNT(*) FROM (
    SELECT PostID FROM Posts GROUP BY PostID HAVING COUNT(DISTINCT BlogName) > 1);

-- rows that reference a blog the registry does not have
SELECT COUNT(*) FROM Posts p
WHERE NOT EXISTS (SELECT 1 FROM Blogs b WHERE b.BlogName = p.BlogName);

SELECT COUNT(*) FROM BlogNames bn
WHERE NOT EXISTS (SELECT 1 FROM Blogs b WHERE b.BlogName = bn.BlogName);

-- space by object, to see where the file actually goes
SELECT name, SUM(pgsize)/1024/1024 AS mb
  FROM dbstat GROUP BY name ORDER BY SUM(pgsize) DESC;

Open the file read-only so an inspection can never disturb a running crawl:

sqlite3 "file:TL.db?mode=ro" ".schema"