SQL Server FileStream - How to acquire files path - sql

Question

I'm working on this application where I need to upload large data files to my SQL Server DB, and I've been using FileStream to do it more efficiently.
I understand the files are stored directly into my system's folders (C:\CryptoDB).
The thing is, I need to manipulate these files (decrypt them) but I haven't been able to recover their filepath.
Doing so, I would be able to manipulate them directly, not having to re-download them via SQL, which is a real waste.
What I've been able to do so far:
My table:
CREATE TABLE [arquivo] (
[idUsuario] INT NOT NULL,
[fileState] INT NOT NULL,
[fileContent] varbinary(max) FILESTREAM,
[fileName] VARCHAR (150) NULL,
[fileSize] VARCHAR (50) NULL,
id UNIQUEIDENTIFIER ROWGUIDCOL NOT NULL UNIQUE,
CONSTRAINT [FK_arquivo_usuario] FOREIGN KEY ([idUsuario]) REFERENCES usuario(id)
);
Insert:
Insert into arquivo(id, idUsuario, fileState, fileContent, fileName, fileSize) Values(
newId(),
1,
5,
(SELECT * FROM OPENROWSET(BULK 'c:\medio.jpeg', SINGLE_BLOB) AS varbinary(max)) ,
'medio.jpeg',
'123'
)
And when I try to recover the filepath:
DECLARE #filePath varchar(max)
SELECT #filePath = fileContent.PathName()
FROM arquivo
PRINT #filepath
The result I get:
\\TEHORT-PC\MSSQLSERVER\v02-A60EC2F8-2B24-11DF-9CC3-AF2E56D89593\CryptoDB\dbo\arquivo\fileContent\31E3697E-0576-4B0F-B0AA-6E046F4116A1\VolumeHint-HarddiskVolume2
Where the file actually is:
C:\CryptoDB\DATA\902a7d8d-c8c1-43b0-8c94-b12319293f42\7febdbd1-02c6-4b00-aa3c-a72bee80ef9c\

Answer 1

SQL query for get physical location of all FILESTREAM data (source)
SELECT t.name AS 'table',
c.name AS 'column',
fg.name AS 'filegroup_name',
dbf.type_desc AS 'type_description',
dbf.physical_name AS 'physical_location'
FROM sys.filegroups fg
INNER JOIN sys.database_files dbf
ON fg.data_space_id = dbf.data_space_id
INNER JOIN sys.tables t
ON fg.data_space_id = t.filestream_data_space_id
INNER JOIN sys.columns c
ON t.object_id = c.object_id
AND c.is_filestream = 1
All FILESTREAM BLOB-fields query result sample
SQL query for get subfolders for FILESTREAM data on server:
(These tables only used within dedicated Administrator connection (DAC)).
SELECT o.name AS [Table], cp.name AS [Column], r.rsguid AS [Rowset GUID], rs.colguid AS [Column GUID] FROM SYS.SYSROWSETS r CROSS APPLY sys.sysrscols rs
JOIN sys.partitions p ON rs.rsid = p.partition_id
JOIN sys.objects o ON o.object_id = p.object_id
JOIN sys.syscolpars cp ON cp.colid = rs.rscolid
WHERE rs.colguid IS NOT NULL AND o.object_id = cp.id AND r.rsguid IS NOT NULL AND r.rowsetid = rs.rsid AND o.name = 'DOCUMENT' and cp.name = 'DIGITAL_FILE';
2.1. Query result:
Table: DOCUMENT
Column: DIGITAL_FILE
Rowset GUID: 0x6AA5E6045794D34D8B1FAC0F49A49B0A
Column GUID: 0xD756E638FB2CC843AE98F489B57F6D7D
Calculating Sub-Path from this guids:
0x6AA5E6045794D34D8B1FAC0F49A49B0A equals this path:
04e6a56a-9457-4dd3-8b1f-ac0f49a49b0a
[reversed 6AA5E604]-[reversed 5794]-[reversed D34D]-[reversed 8B1F]-[original AC0F49A49B0A]
0xD756E638FB2CC843AE98F489B57F6D7D equals this path:
38e656d7-2cfb-43c8-ae98-f489b57f6d7d (rules in previus guid parsing)
2.2 Result calculated full path for FILESTREAM storage:
i:\SQL Base posc_astrachan FileStreams\GTMK\GTM_FILE_STREAM\04e6a56a-9457-4dd3-8b1f-ac0f49a49b0a\38e656d7-2cfb-43c8-ae98-f489b57f6d7d
Get original filename for BLOB-value in NTFS-folder.
3.1. Stored procedure for query advanced SQL Server Page info
SET ANSI_NULLS ON
SET QUOTED_IDENTIFIER ON
GO
CREATE PROCEDURE [dbo].[procDBCC_PAGE]
#db_name varchar (500),
#filenum INT,
#pagenum INT
AS
BEGIN
SET NOCOUNT ON
DBCC TRACEON (3604);
DBCC PAGE (#db_name, #filenum, #pagenum, 3) WITH TABLERESULTS;
SET NOCOUNT OFF
END
3.2. Stored procedure for query original filename for FILESTREAM'ed BLOB-field of table
SET ANSI_NULLS ON
SET QUOTED_IDENTIFIER ON
GO
CREATE PROCEDURE [dbo].[procFindLogSequenceNumber]
-- #TableName varchar (500),
#instanceS varchar (19), -- key value for filed INSTANCE_S
#tableName varchar(500), -- DOCUMENT
#keyFieldName varchar(500), -- INSTANCE_S
#LogSequenceNumber varchar (500) OUTPUT
AS
SET NOCOUNT ON
DECLARE #db_name varchar (500)
DECLARE #filenum INT
DECLARE #pagenum INT
DECLARE #slotnum INT
DECLARE #rid varchar (100)
DECLARE #ridDotted varchar (100)
DECLARE #parent_object varchar (500)
DECLARE #sql nvarchar(2000)
DECLARE #sqlTable Table(physloc varchar(100))
DECLARE #DBCC_PAGE_Output Table ([ParentObject] varchar (MAX), [Object] varchar (MAX), [Field] varchar (MAX), [VALUE] varchar (MAX))
SET #db_name = db_name()
SET #sql = 'SELECT top 1 sys.fn_PhysLocFormatter (%%physloc%%) AS [PhysicalRID] FROM '+#tableName+' WHERE '
+#keyFieldName+' = '''+#instanceS+''''
INSERT #sqlTable (physloc)
EXECUTE sp_executesql #sql
SET #rid = (select top 1 physloc from #sqlTable)
if #rid is NULL
BEGIN
RETURN -1;
END
-- parse (#rid): (1:1172779:6) 1-#filenum, 2- #pagenum, 3- #slotnum
SET #ridDotted = Replace(#rid, ':', '.');
SET #ridDotted = Replace(#ridDotted, '(', '');
SET #ridDotted = Replace(#ridDotted, ')', '');
SET #filenum = (SELECT Parsename(#ridDotted, 3))
SET #pagenum = (SELECT Parsename(#ridDotted, 2))
SET #slotnum = (SELECT Parsename(#ridDotted, 1))
INSERT #DBCC_PAGE_Output ([ParentObject], [Object], [Field], [VALUE])
EXECUTE procDBCC_PAGE #db_name, #filenum , #pagenum
SET #parent_object = (SELECT TOP 1 [ParentObject] FROM #DBCC_PAGE_Output WHERE [Field] = 'INSTANCE_S'
AND [VALUE] = #instanceS)
--CreateLSN field Only
SET #LogSequenceNumber = (SELECT [VALUE] FROM #DBCC_PAGE_Output WHERE
[ParentObject] = #parent_object AND
[Field] = 'CreateLSN'
)
if #LogSequenceNumber is NULL
BEGIN
RETURN -1;
END
-- result 0006c050:00000120:0090 (442448:288:144)
-- clear (...)
SET #LogSequenceNumber = Replace(#LogSequenceNumber, ' ', '.');
SET #LogSequenceNumber = (SELECT Parsename(#LogSequenceNumber, 2))
--replace ":" to "-"
SET #LogSequenceNumber = Replace(#LogSequenceNumber, ':', '-');
SET NOCOUNT OFF
3.3. Sample query for stored procedure for get filename on NTFS folder for BLOB:
declare #filestreamFileName varchar(500);
exec procFindLogSequenceNumber 'ZW_NU9hGZ0CKoSXYAoc', 'DOCUMENT', 'INSTANCE_S', #filestreamFileName OUTPUT
select #filestreamFileName
3.4. Result (original file name in NTFS folder):
0003137a-00001244-00d0
3.5. Result full path:
i:\SQL Base posc_astrachan FileStreams\GTMK\GTM_FILE_STREAM\04e6a56a-9457-4dd3-8b1f-ac0f49a49b0a\38e656d7-2cfb-43c8-ae98-f489b57f6d7d\0003137a-00001244-00d0

Answer 2

The path you get is correct, you are supposed to get the network share path not the local path and use SqlFileStream to open a stream.
https://msdn.microsoft.com/en-us/library/system.data.sqltypes.sqlfilestream%28v=vs.110%29.aspx
You could also get a file handle using OpenSqlFilestream method, i.e. to use it in Windows API.
https://msdn.microsoft.com/en-us/library/bb933972.aspx

Answer 3

Alexander's answer is great, saved me a LOT of trouble correlating page/slot numbers to actual LSNs. In my case using SQL Server 2008 R2 I had to make some adjustments to his SP to get it working, those were:
The SPs first parameter:
#instanceS varchar (100), -- key value for filed INSTANCE_S
In my case this is a uniqueidentifier so I needed a bigger varchar, original value was 19.
When querying DBCC PAGE Output:
SET #parent_object = (SELECT TOP 1 [ParentObject] FROM
#DBCC_PAGE_Output WHERE [Field] = #keyFieldName AND [VALUE] =
#instanceS)
It originaly stated "[Field] = 'INSTANCE_S'" which apparently hardcoded a value that worked for OP but not for me. It needed to match the name of the FILESTREAM table's key field.
Also to clarify the SPs input params a bit:
#instanceS = The actual column value which identifies the row. Will
this always match the column set as the table's "RowGuid"?
#tableName = Pretty clear. The FILESTREAM table's name.
#keyFieldName = The name of the table's key column. Should be the
source column from where #instanceS was taken.