Re: [sqlite] [EXTERNAL] Inconsistency of CREATE/DROP TABLE with attached DBs

Addendum: Note that CREATE TABLE has an optional schema name which must be one of - main (which is the default) - temp (which is the same as specifying TEMP or TEMPORARY between CREATE and TABLE) - the name of an attached database See https://sqlite.org/lang_createtable.html

Re: [sqlite] [EXTERNAL] Inconsistency of CREATE/DROP TABLE with attached DBs

I don't see any inconsistency here. 1) implicit attach of a.sqlite as main and create a.t1 2) implcit attach b.sqlite as main, attach a.sqlite as a and create (main).t1 (in b.sqlite) 3) implicit attach c.sqlite as main, attaxh a.sqlite as a and drop the only table named t1 from a c.sqlite

[sqlite] Inconsistency of CREATE/DROP TABLE with attached DBs

I ran into an inconsistency? between CREATE and DROP TABLE today: # Pipe this example through sed -e 's/^ *//' before running through # a shell # # 1. Set up table a.t1 rm -f a.sqlite b.sqlite c.sqlite cat

Re: [sqlite] Shell Feature Request: spaces before dot commands

On Wed Mar 04, 2020 at 12:44:09PM +, Simon Slavin wrote: > On 4 Mar 2020, at 12:19pm, no...@null.net wrote: > > >I like to sometimes indent a block of SQL and change > > settings or run dot commands within a transaction, e.g.: > > > >BEGIN > >.mode csv > >.import ...

Re: [sqlite] Getting all changes within a begin; end; transaction

Simon Slavin, on Wednesday, March 4, 2020 10:47 AM, wrote... > > On 4 Mar 2020, at 3:28pm, Jose Isaias Cabrera wrote: > > > The reason why I know is that if I have 238 INSERTS, but I have a > constraint , there should be 238 INSERTs the first time I run a set of SQL, > but if I run the same SQL

Re: [sqlite] Getting all changes within a begin; end; transaction

On 4 Mar 2020, at 3:28pm, Jose Isaias Cabrera wrote: > The reason why I know is that if I have 238 INSERTS, but I have a constraint > , there should be 238 INSERTs the first time I run a set of SQL, but if I run > the same SQL again, there should not be any records INSERTED, and thus, the >

Re: [sqlite] Getting all changes within a begin; end; transaction

On 4 Mar 2020, at 3:28pm, Jose Isaias Cabrera wrote: > However, I have found that it does not actually provide the **ACTUAL** > changes, but a count of the possible changes. Hmm. I understand you. Does this do something more like what you need ?

Re: [sqlite] Getting all changes within a begin; end; transaction

Simon Slavin, on Wednesday, March 4, 2020 09:42 AM, wrote... > > On 4 Mar 2020, at 2:37pm, Jose Isaias Cabrera wrote: > > > Is there a way to know all the changes that may have happened within > the full BEGIN and END? Thanks. > > Use this function > >

Re: [sqlite] Getting all changes within a begin; end; transaction

On 4 Mar 2020, at 2:37pm, Jose Isaias Cabrera wrote: > Is there a way to know all the changes that may have happened within the full > BEGIN and END? Thanks. Use this function before and after your block, and subtract one from another.

[sqlite] Getting all changes within a begin; end; transaction

Greetings. Imagine this SQL, BEGIN TRANSACTION; ... changes to records ... END; When I execute "int result = sqlite3_changes(database);" after that SQL execution, I always get 1. I think that it is because it is only providing the result of the last statement that was successful within the

Re: [sqlite] Shell Feature Request: spaces before dot commands

On 4 Mar 2020, at 12:19pm, no...@null.net wrote: >I like to sometimes indent a block of SQL and change > settings or run dot commands within a transaction, e.g.: > >BEGIN >.mode csv >.import ... >COMMIT; Dot commands are not SQL commands, they are instructions to the

[sqlite] Shell Feature Request: spaces before dot commands

[SQLite version 3.22.0 2017-11-27 17:56:14] The SQLite shell only recognizes .dot commands without leading spaces. For clarity I like to sometimes indent a block of SQL and change settings or run dot commands within a transaction, e.g.: BEGIN .mode csv .import ... COMMIT;