[jira] [Comment Edited] (IGNITE-10784) SQL: Create a view with list of existing tables
[
https://issues.apache.org/jira/browse/IGNITE-10784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16744875#comment-16744875
]
Pavel Kuznetsov edited comment on IGNITE-10784 at 1/17/19 10:30 AM:
[~vozerov], I've updated PR according your comments.
was (Author: pkouznet):
[~vozerov] Updated according your comments.
> SQL: Create a view with list of existing tables
> ---
>
> Key: IGNITE-10784
> URL: https://issues.apache.org/jira/browse/IGNITE-10784
> Project: Ignite
> Issue Type: Task
> Components: sql
>Reporter: Vladimir Ozerov
>Assignee: Pavel Kuznetsov
>Priority: Major
> Fix For: 2.8
>
>
> We need to create a system view of currently available SQL tables.
> Minimal required information:
> 1) Schema name
> 2) Table name
> 3) Owning cache name
> 4) Owning cache ID
> Other info to consider:
> 1) Affinity column name
> 2) Key/value aliases
> 3) Key/value type names
> 4) Analyse other vendors (e.g. MySQL, Postgresql) and see if any other useful
> information could be exposed (taking in count that a lot of engine properties
> are already exposed through {{CACHES}} view)
> Starting point: {{SqlSystemView}}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Comment Edited] (IGNITE-10784) SQL: Create a view with list of existing tables
[
https://issues.apache.org/jira/browse/IGNITE-10784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16740199#comment-16740199
]
Vladimir Ozerov edited comment on IGNITE-10784 at 1/11/19 9:22 AM:
---
[~pkouznet], my comments:
# It is not clear what is the reason of adding locking to
{{GridH2Table.getColumns}} method?
# {{SqlSystemViewTables}} - no need to define separate constants for every
field name.
# Naming: {{mngr}} -> {{mgr}}
# Some column renames are needed: {{OWNING_CACHE_NAME}} -> {{CACHE_NAME}},
{{OWNING_CACHE_ID}} -> {{CACHE_ID}}, {{TABLE_SCHEMA}} -> {{SCHEMA_NAME}}
# Why do we need to use {{ctx.cache().publicCacheNames()}}, while
{{SchemaManager}} already has all tables?
# {{SqlSystemViewTables.getRowCount}} - same thing. Why don't just call
{{SchemaManager.dataTables.size()}}?
# Key and value aliases should be {{null}} if they are not set
# There is no need to iterate over columns to find key/val alias. Instead, you
can get key or val alias ID from {{GridH2RowDescriptor}} and dereference that
column directly.
# Test coverage needs to be improved a bit - we need more queries with
conditions on different columns.
was (Author: vozerov):
[~pkouznet], my comments:
# It is not clear what is the reason of adding locking to
{{GridH2Table.getColumns}} method?
# {{SqlSystemViewTables}} - no need to define separate constants for every
field name.
# Naming: {{mngr}} -> {{mgr}}
# Some column renames are needed: {{OWNING_CACHE_NAME}} -> {{CACHE_NAME}},
{{OWNING_CACHE_ID}} -> {{CACHE_ID}}, {{TABLE_SCHEMA}} -> {{SCHEMA_NAME}}
# Why do we need to use {{ctx.cache().publicCacheNames()}}, while
{{SchemaManager}} already has all tables?
# {{SqlSystemViewTables.getRowCount}} - same things. We don't just call
{{SchemaManager.dataTables.size()}}?
# Key and value aliases should be {{null}} if they are not set
# There is no need to iterate over columns to find key/val alias. Instead, you
can get key or val alias ID from {{GridH2RowDescriptor}} and dereference that
column directly.
# Test coverage needs to be improved a bit - we need more queries with
conditions on different columns.
> SQL: Create a view with list of existing tables
> ---
>
> Key: IGNITE-10784
> URL: https://issues.apache.org/jira/browse/IGNITE-10784
> Project: Ignite
> Issue Type: Task
> Components: sql
>Reporter: Vladimir Ozerov
>Assignee: Pavel Kuznetsov
>Priority: Major
> Fix For: 2.8
>
>
> We need to create a system view of currently available SQL tables.
> Minimal required information:
> 1) Schema name
> 2) Table name
> 3) Owning cache name
> 4) Owning cache ID
> Other info to consider:
> 1) Affinity column name
> 2) Key/value aliases
> 3) Key/value type names
> 4) Analyse other vendors (e.g. MySQL, Postgresql) and see if any other useful
> information could be exposed (taking in count that a lot of engine properties
> are already exposed through {{CACHES}} view)
> Starting point: {{SqlSystemView}}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Comment Edited] (IGNITE-10784) SQL: Create a view with list of existing tables
[
https://issues.apache.org/jira/browse/IGNITE-10784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16728692#comment-16728692
]
Pavel Kuznetsov edited comment on IGNITE-10784 at 12/25/18 12:35 PM:
-
Lets add also "constant" fields : CATALOG_NAME ("IGNITE"), TABLE_TYPE ("TABLE")
It is also good to have TABLE_ROWS field, I should think, if we can do this
without scan query every time (size of pk index?).
was (Author: pkouznet):
Lets add also "constant" fields : CATALOG_NAME ("IGNITE"), TABLE_TYPE ("TABLE")
It is also good to have TABLE_ROWS field, should think, if we can do without
scan (size of pk index?).
> SQL: Create a view with list of existing tables
> ---
>
> Key: IGNITE-10784
> URL: https://issues.apache.org/jira/browse/IGNITE-10784
> Project: Ignite
> Issue Type: Task
> Components: sql
>Reporter: Vladimir Ozerov
>Assignee: Pavel Kuznetsov
>Priority: Major
> Fix For: 2.8
>
>
> We need to create a system view of currently available SQL tables.
> Minimal required information:
> 1) Schema name
> 2) Table name
> 3) Owning cache name
> 4) Owning cache ID
> Other info to consider:
> 1) Affinity column name
> 2) Key/value aliases
> 3) Key/value type names
> 4) Analyse other vendors (e.g. MySQL, Postgresql) and see if any other useful
> information could be exposed (taking in count that a lot of engine properties
> are already exposed through {{CACHES}} view)
> Starting point: {{SqlSystemView}}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Comment Edited] (IGNITE-10784) SQL: Create a view with list of existing tables
[
https://issues.apache.org/jira/browse/IGNITE-10784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16728651#comment-16728651
]
Pavel Kuznetsov edited comment on IGNITE-10784 at 12/25/18 12:33 PM:
-
maybe we should duplicate atomic mode or introduce new column "mvcc enabled".
Upd: I mean column with boolean type, which is true if owning cache has
atomicity mode == TRANSACTIONAL_SNAPSHOT
was (Author: pkouznet):
maybe we should duplicate atomic mode or introduce new column "mvcc enabled".
> SQL: Create a view with list of existing tables
> ---
>
> Key: IGNITE-10784
> URL: https://issues.apache.org/jira/browse/IGNITE-10784
> Project: Ignite
> Issue Type: Task
> Components: sql
>Reporter: Vladimir Ozerov
>Assignee: Pavel Kuznetsov
>Priority: Major
> Fix For: 2.8
>
>
> We need to create a system view of currently available SQL tables.
> Minimal required information:
> 1) Schema name
> 2) Table name
> 3) Owning cache name
> 4) Owning cache ID
> Other info to consider:
> 1) Affinity column name
> 2) Key/value aliases
> 3) Key/value type names
> 4) Analyse other vendors (e.g. MySQL, Postgresql) and see if any other useful
> information could be exposed (taking in count that a lot of engine properties
> are already exposed through {{CACHES}} view)
> Starting point: {{SqlSystemView}}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
[jira] [Comment Edited] (IGNITE-10784) SQL: Create a view with list of existing tables
[
https://issues.apache.org/jira/browse/IGNITE-10784?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel=16728522#comment-16728522
]
Pavel Kuznetsov edited comment on IGNITE-10784 at 12/25/18 9:55 AM:
MySQL has approximate table size.
was (Author: pkouznet):
MySQL has approximate table table size.
> SQL: Create a view with list of existing tables
> ---
>
> Key: IGNITE-10784
> URL: https://issues.apache.org/jira/browse/IGNITE-10784
> Project: Ignite
> Issue Type: Task
> Components: sql
>Reporter: Vladimir Ozerov
>Assignee: Pavel Kuznetsov
>Priority: Major
> Fix For: 2.8
>
>
> We need to create a system view of currently available SQL tables.
> Minimal required information:
> 1) Schema name
> 2) Table name
> 3) Owning cache name
> 4) Owning cache ID
> Other info to consider:
> 1) Affinity column name
> 2) Key/value aliases
> 3) Key/value type names
> 4) Analyse other vendors (e.g. MySQL, Postgresql) and see if any other useful
> information could be exposed (taking in count that a lot of engine properties
> are already exposed through {{CACHES}} view)
> Starting point: {{SqlSystemView}}
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
