From this database, I want to select two users, 321 and 102 where BOOK_TYPE is 1 or 2 and compare their BOOK_NR, if they coincide then save BOOK_COUNT only from USER_ID 102.
SqlCommand command = new SqlCommand("SELECT BOOK_NR, BOOK_COUNT, USER_ID, BOOK_TYPE " +
"FROM BOOKS " +
"WHERE (BOOK_TYPE = '1' OR" + "BOOK_TYPE = '2')" +
"AND MONTH(DATE) = '" + DateTime.Today.Month + "'" +
"GROUP BY BOOK_NR, BOOK_COUNT, USER_ID, BOOK_TYPE ", BookConn);
SqlDataReader reader = command.ExecuteReader();
while (reader.Read())
{
string BookNr = (string)reader[0];
int Count = (int)reader[1];
int User = (int)reader[2];
int Type = (int)reader[3];
if( " HERE I NEED HELP " )
{
" AND HERE :) "
}
}
reader.Close();BookConn.Close();
My only solution is as follows
List<string> User321List = new List<string>();
List<string> User102List = new List<string>();
if(User == 321 && Type == 1){ User321List.Add(BookNr.Trim());}
if(User == 102 && Type == 2){ User102List.Add(BookNr.Trim()+"\t"+Count);}
and then...
int count = 0;
foreach (string x in User321List)
{
foreach (string y in User102List)
{
List<string> part = y.Split('\t').Select(p => p.Trim()).ToList();
var z = string.Compare(part[0], x);
if (z == 0)
{
count += int.Parse(part[1]);
}
}
}
but it takes a lot of time to get result..
if someone figured something out of my nightmare please help and sorry for my bad english...
I'm going to focus on answering just the first part:
I want to select two users, 321 and 102 where BOOK_TYPE is 1 or 2 and compare their BOOK_NR, if they coincide then save BOOK_COUNT only from USER_ID 102.
You want to join the table to itself:
SELECT b1.BOOK_NR, b1.BOOK_COUNT, b1.USER_ID, b1.BOOK_TYPE
FROM BOOKS b1
INNER JOIN BOOKS b2 ON b2.BOOK_TYPE IN ('1','2')
AND b2.User_ID = 321
AND b1.BOOK_NR = b2.BOOK_NR
WHERE b1.User_ID = 102 AND b1.BOOK_TYPE IN ('1','2')
AND MONTH(b1.DATE) = MONTH(current_timestamp)
AND MONTH(b2.DATE) = MONTH(current_timestamp)
Since one of your concerns was performance -- that the query was too slow -- I also need to ask whether the BOOK_TYPE column is a string or a number. If it's a number column, you should omit the single quotes on the values: for example, just use IN (1,2) instead of IN ('1','2'). This will help performance by avoiding potential per-row conversions and by potentially better matching to indexes.
If your query is slow, consider splitting the operation in two. Create a SQL Table function so you don't repeat the SQL, passing the users, type and month in as parameters.
Do not do an IN (as suggested in comment). Keep in mind that an OR in the WHERE clause will result in no indexes being used.
Join the two instances of the table function and let sql do the matching for you.
e.g.
Select * from dbo.UserList('321',1,12) as u321
inner join dbo.UserList('102',2,12) as u102
on u321.field = u102.field
BTW... I covered quite a few advanced topics, but the idea is to let SQL do what it does best: JOINS
As a matter of interest, if your sql table function returns many rows, and you ever use them in a sp, remember to use the option (recompile) clause ...
Related
I am new to C# (has experience in other languages before, C++, SQL AutoIT). I have a datatable with 10 columns
Name, MemberNoA, MemberNoB, DriverLicense, MobileNo, Address1, Address2, Address3, ProgramJoinned, Remark
The datatable has around 17,000 rows, what I want to do is, if the same person's records appear more than 2 times in the datatable, put a description in a remark field.
4 criteria to define "same person", any one criteria match will treat as "same person"
i Name + MemberNoA
ii Name + MemberNoB
iii Name + DriverLicense
iv Name + MobileNo
i.e. if there are 3 records with same Name and same MemberNoA, need to put description into remark field of these 3 records.
I work out result set from the above 4 criteria like this:
var resultCriteria1 = from gpr in dt.AsEnumerable()
group gpr by new {
Fld1 = gpr.Field < string > ("Name"),
Fld2 = gpr.Field < string > ("MemberNoA")
}
into grpp
where grpp.Count() > 2
select new {
Name = grpp.Key.Fld1,
MemA = grpp.Key.Fld2,
Cnt = grpp.Count()
};
after that, I loop thru all rows in dt and for each row, loop thru all result set in 4 criteria:
for (int i = 1; i < dt.Rows.Count; i++) {
foreach(var item in resultCriteria1) {
if ((item.Nam == s trName) && (item.MemA == M emberNoA)) {
dt.Rows[i].ItemArray[9] = d t.Rows[i].ItemArray[9] + "Criteria 1 match\r\n";
}
}
}
The program work but run very slow! Is there any method like simple sql statement:
update table1 where
table1.name = table2.name and
table1.MemberNoA = table2.MemberNo2
set table1.Remark = "Criteria 1 match\r\n"
Is there any way to do this in C# or any way to optimize it ? Thanks.
Regds
LAM Chi-fung
The problem is that you are making a cartesian join between the grouped results and the original datatable, without using any performant data structures such as a dictionary or hashset.
But you don't actually need to join it at all, the grouped results can actually hold the relevant data rows directly.
The following code should be performant enough
var grouped =
from gpr in dt.Rows.Cast<DataRow>()
group gpr by (
Fld1: (string)gpr["Name"],
Fld2: (string)gpr["MemberNoA"]
)
into grpp
where grpp.Count() > 2
select grpp;
foreach (var grp in grouped)
foreach (var row in grp)
row["Remark"] += "Criteria 1 match\r\n";
What you can do is to use hashtable, order your data, and then iterate comparing current row data with previous using a cursor. This should give you Log(n) time complexity.
I have two different tables from a DataBase named "empleados" and "fichajes". empleados has the employees data and fichajes has the date and time from when they started working.
I want to get the total work time done by a specific employee between two dates, lets say from 20th to 29th.
I have this query which I use with Dapper on C#:
SELECT CONCAT(e.nombre, " " ,e.apellido) as fullName, tfichajes.total
FROM empleados e
INNER JOIN (
SELECT f1.nif,
SEC_TO_TIME(SUM(TIME_TO_SEC(f1.fechasalida) - TIME_TO_SEC(f1.fechaentrada))) AS total
FROM fichajes f1
WHERE f1.fechasalida <= '2019-04-29'
and f1.fechaentrada >= '2019-04-20'
GROUP BY f1.nif
) AS tfichajes
ON e.nif = tfichajes.nif
WHERE e.nif = '33333333P'
This works just fine, but I was wondering if it was possible to make it simpler.
This is the code I have in my program:
public static List<string> CalculaTotalHoras(string nif, DateTime fechaEntrada, DateTime fechaSalida)
{
var dbCon = DBConnection.Instancia();
if (dbCon.Conectado())
{
string format = "yyyy-MM-dd";
List<string> result = new List<string>();
using (IDbConnection conexion = dbCon.Conexion)
{
var output = conexion.Query($"SELECT CONCAT(e.nombre, \" \" ,e.apellido) as fullName, tfichajes.total " +
$"FROM empleados e INNER JOIN (SELECT f1.nif, SEC_TO_TIME(SUM(TIME_TO_SEC(f1.fechasalida) - TIME_TO_SEC(f1.fechaentrada))) AS total " +
$"FROM fichajes f1 where f1.fechasalida <= '{fechaSalida.ToString(format)}' and f1.fechaentrada >= '{fechaEntrada.ToString(format)}' GROUP BY f1.nif) " +
$"as tfichajes ON e.nif = tfichajes.nif where e.nif = '{ nif }';").ToList();
var i = 0;
foreach (IDictionary<string, object> row in output)
{
foreach (var pair in row)
{
if (i == 0)
{
result.Add(pair.Value.ToString());
i++;
}
else
{
result.Add(pair.Value.ToString());
}
}
}
return result;
}
}
else return null;
}
If you have problems with the readability of the code here you have a gyazo.
Workingtime fichajes table and employees empleados table.
With that exact query, the expected results are Alvaro Arguelles 00:05:00, and in the code, I want to get Alvaro Arguelles and 00:05:00 separated in the result List.
Did I made it much harder than it actually is?
Your query can be significantly simplified. Your inner query is selecting ALL employees working in the date in question, I would start with the INNER query as the main FROM table getting the data for the one employee. THEN join to the employee to grab the name.
select
CONCAT(e.nombre, " " ,e.apellido) as fullName,
OneEmp.Total
from
( select
f1.nif,
SEC_TO_TIME( SUM(TIME_TO_SEC(f1.fechasalida)
- TIME_TO_SEC(f1.fechaentrada))) AS total
from
fichajes f1
where
f1.nif = '33333333P'
AND f1.fechasalida >= '2019-04-20'
AND f1.fechasalida <= '2019-04-29' ) OneEmp
JOIN empleados e
on OneEmp.nif = e.nif
Since your query was summarizing per the one employee, it would only return a single row anyhow. So the above query will return 1 row, 2 columns... name and total. Should be able to read those two values back directly.
But as others commented, PARAMETERIZE your query calls.
I have this Method# 1 query below that is parameterized using dapper, problem is the query times out with this approach even after waiting 30sec and normally it takes max of 1 sec on SSMS with plain sql.
However Method # 2 query actually works where the query is built on the server side instead of parameterized one. One thing i have noticed is, it might have something to do with filter for FirstName and LastName, I have single Quote on Method #2 for those filter but not for Method #1.
What is wrong with Method # 1 ?
Method # 1
string query = "SELECT *
FROM dbo.Customer c
WHERE c.MainCustomerId = #CustomerId
AND (#IgnoreCustomerId = 1 OR c.CustomerID = #FilterCustomerId)
AND (#IgnoreFirstName = 1 OR c.FirstName = #FilterFirstName)
AND (#IgnoreLastName = 1 OR c.LastName = #FilterLastName)
AND (#IgnoreMemberStatus = 1 OR c.CustomerStatusID = #FilterMemberStatus)
AND (#IgnoreMemberType = 1 OR c.CustomerTypeID = #FilterMemberType)
AND (#IgnoreRank = 1 OR c.RankID = #FilterRank)
ORDER BY c.CustomerId
OFFSET #OffSet ROWS
FETCH NEXT 50 ROWS ONLY";
_procExecutor.ExecuteSqlAsync<Report>(query, new
{
CustomerId = customerId,
IgnoreCustomerId = ignoreCustomerId,
FilterCustomerId = filter.CustomerID,
IgnoreFirstName = ignoreFirstName,
FilterFirstName = filter.FirstName,
IgnoreLastName = ignoreLastName,
FilterLastName = filter.LastName,
IgnoreMemberStatus = ignoreMemberStatus,
FilterMemberStatus = Convert.ToInt32(filter.MemberStatus),
IgnoreMemberType = ignoreMemberType,
FilterMemberType = Convert.ToInt32(filter.MemberType),
IgnoreRank = ignoreRank,
FilterRank = Convert.ToInt32(filter.Rank),
OffSet = (page - 1) * 50
});
Method # 2
string queryThatWorks =
"SELECT *
FROM dbo.Customer c
WHERE c.MainCustomerId = #CustomerId
AND ({1} = 1 OR c.CustomerID = {2})
AND ({3} = 1 OR c.FirstName = '{4}')
AND ({5}= 1 OR c.LastName = '{6}')
AND ({7} = 1 OR c.CustomerStatusID = {8})
AND ({9} = 1 OR c.CustomerTypeID = {10})
AND ({11} = 1 OR c.RankID = {12})
ORDER BY c.CustomerId
OFFSET {13} ROWS
FETCH NEXT 50 ROWS ONLY";
_procExecutor.ExecuteSqlAsync<Report>(string.Format(queryThatWorks,
customerId,
ignoreCustomerId,
filter.CustomerID,
ignoreFirstName,
filter.FirstName,
ignoreLastName,
filter.LastName,
ignoreMemberStatus,
Convert.ToInt32(filter.MemberStatus),
ignoreMemberType,
Convert.ToInt32(filter.MemberType),
ignoreRank,
Convert.ToInt32(filter.Rank),
(page - 1) * 50
), null);
I've seen this countless times before.
I'm willing to bet that your columns are varChar, but Dapper is sending in your parameters as nVarChar. When that happens, SQL Server has to run a conversion on the value stored in each and every row. Besides being really slow, this prevents you from using indexes.
See "Ansi Strings and varchar" in https://github.com/StackExchange/dapper-dot-net
Database structure sample:
Department Table
-DepartmentID
Facility Table
-Facility ID
-DepartmentID (FK)
-Block
-Level
-Name
-etc
I am trying to select from database from the EF using two where clause. I am not sure what went wrong at the where clause. I am stuck at it. Please help and advice. I have googled on the internet but cannot find the solution to it.
string departmentID = "SIT";
string block = "L";
string level = "4";
string name = "L.425";
using (var db = new KioskContext())
{
var facilitys = from f in db.Facilitys
Where clause to select departmentID where equals to SIT and also where any block or level or name contains any alphabets. Please advice how should i write the statement with two where clause. Thank You!
where f.Department.DepartmentID == departmentID
&& (f.Block.Contains("%" + block + "%") || f.Level.Contains("%" + level + "%")
|| f.Name.Contains("%" + name + "%"))
Remaining of the query statement to select all the facilities
orderby f.FacilityID
select new
{
f.FacilityID,
f.DepartmentID,
f.Description,
f.Block,
f.Level,
f.Name,
f.OpenHours,
f.CloseHours,
f.MaxBkTime,
f.MaxBkUnits,
f.MinBkTime,
f.MinBkUnits
};
foreach (var fac in facilitys)
{
FacObject facobject = new FacObject(fac.FacilityID, fac.DepartmentID, fac.Description, fac.Block, fac.Level,
fac.Name, fac.OpenHours, fac.CloseHours, fac.MaxBkTime, fac.MaxBkUnits, fac.MinBkTime, fac.MinBkUnits);
sqlFacList.Add(facobject);
}
}
Remove the "%" from the various Contains clauses, they are SQL cruft you do not need.
where f.Department.DepartmentID == departmentID
&& (f.Block.Contains(block)
|| f.Level.Contains(level)
|| f.Name.Contains(name ))
Remember LINQ is not just for SQL!
Query Detail Table
Main Table
1 .PK
Detail Table
1. PK
2. Detail description.
If mysql i can group the main table PK for duplicate issue but how it implement in sql server ?Since sql server required something like aggregate
E.g
(vehicle) Main Table
Car
bikes
van
(vehicleItem) Detail Table
Car->item a,item b
bikes-> item a,item c..
van->item b ,item c
I want to filter in a query such as i want to filter item a only?Seem quite inefficient to call in sub loop.Even put all detail item it in a column consider as incorrect but when search item id(number) might be same number unless the item id are guid and distinct different.
** the main purpose is to output list not to sum crosstab :) ya.. grouping cool if you want to sum up.
var sql = "select * from vehicle ";
var command = new SqlCommand(connection,sql);
try {
var reader = command.ExecuteReader();
if (reader.HasRows) {
while (reader.Read())
if (!string.IsNullOrEmpty(Request.QueryString["ItemIdValue[]"])){
var d = Request.QueryString.GetValues("ItemIdValue[]");
if (d != null){
if (d[0].Contains("all")){
if (GetItemExist(" AND vehicleItem.itemId in (SELECT itemId from item ) ")){ }
}else{
if (GetItemExist(" AND vehicleItem.itemId IN ( " + itemFilter.Remove(itemFilter.Length - 1) + " ) ")){ }
}
}
}
}
}
}
At last thinking back, why not select main table and in subquery..
var sql = "select * from vehicle ";
if (!string.IsNullOrEmpty(Request.QueryString["itemIdValue[]"])){
var d = Request.QueryString.GetValues("itemIdValue[]");
if (d != null){
if (d[0].Contains("all")){
sql = sql + #"
AND vehicleId IN (
SELECT vehicleId
FROM vehicle
JOIN vehicleItem
ON vehicle.vehicleId = vehicleItem.vehicleId
WHERE vehicleItem.itemId in
(SELECT itemId from item)
) ";
} else{
for (var e = 0; e < d.Length; e++){
itemFilter += d[e] + ",";
_fieldVariable.Add("itemIdValue[]");
_valueVariable.Add(d[e]);
}
sql = sql + #"
AND vehicleId IN (
SELECT vehicleId
FROM vehicle
JOIN vehicleItem
ON vehicle.vehicleId = vehicleItem.vehicleId
WHERE vehicleItem.itemId in ( " + itemFilter.Remove(itemFilter.Length - 1) + " ) ) ";
}
}
}
When joining two tables t1 and t2, a result set containing all matching rows from table t1 and table t2 is created. When you only want one resulting row per t1's row, it is up to you to filter the join accordingly.
There are several possible ways to do that:
adapt the join's ON clause
group accordingly
specify a WHERE clause
As you wanted to use grouping in your request, you could use a query like this:
SELECT m.PK, min(d.Detail)
FROM Main m
LEFT JOIN Detail d on d.PK=m.pk
WHERE d.Detail = 'item a'
GROUP by m.PK
That way you will get one row per PK with the minimum value in detail. You can use other aggregate functions too. See https://msdn.microsoft.com/en-us/library/ms173454.aspx.
Edit: Added WHERE clause to filter by "item a".