I'm trying to use ThinkingSphinx to return records that have a start date within a range OR an end date within the same range, basically any record that starts or ends within this range.
To do this, I am using a computed attribute and sphinx_select as per the documentation in combination with what this post suggests for date ranges, as follows (assume there are two records, record_a starts outside the range, but ends within the range and record_b starts and ends within the range):
with_display = "*, IF(start_at >= #{range_start.to_i}, 1, 0) + " +
"IF(start_at <= #{range_end.to_i}, 1, 0) + " +
"IF(end_at >= #{range_start.to_i}, 10, 0) + " +
"IF(end_at <= #{range_end.to_i}, 10, 0) AS display"
{
sphinx_select: with_display,
with: {'display' => [2, 20, 22]},
}
=> [record_b]
However, if I only use the start_at conditions, I get one record, and if I use only the end_at conditions, it returns both records.
with_display = "*, IF(start_at >= #{range_start.to_i}, 1, 0) + " +
"IF(start_at <= #{range_end.to_i}, 1, 0) AS display"
=> [record_b]
with_display = "*, IF(end_at >= #{range_start.to_i}, 10, 0) + " +
"IF(end_at <= #{range_end.to_i}, 10, 0) AS display"
=> [record_a, record_b]
If I'm understanding this correctly, having all four conditions, should result in both record_a and record_b being returned, since record_a should have a display value of 20, while record_b should have a display value of 22.
Am I missing something?
I just realized my math was wrong, given the cases I want to handle:
record_a will have a display of 21
record_b will have a display of 22
What I needed to do was change my array to:
{
sphinx_select: with_display,
with: {'display' => [12, 21, 22]},
}
in order the handle the cases of records that end within the range (21), records that start within range (12), and records than start and end within the range (22)
Related
I keep track of trading data in a private spreadsheet and, once a trade is closed, I push a button to run some AppsScript to copy the data across to a publicly visible spreadsheet. It all works well except the following point. I put the dates and times a trade is opened in two different columns and the same with the dates and times trades are closed. When I copy the dates across to the target spreadsheet, instead of "10 Jul 2022" it is showing as "10/07/2022 02:00:00". When I sort the spreadsheet on these date and time columns, this can often produce unwated results. What I want is for the date column to just display as "10 Jul 2022" and for that data to recognized as a date for sorting purposes. Is that possible?
function updatePublicSheet(){
var coinBought;
var dateBought;
var timeBought;
var priceBought;
var dateSold;
var timeSold;
var priceSold;
var sLrIndex;
var tLrIndex;
var sss = SpreadsheetApp.openByUrl("https://docs.google.com/spreadsheets/d/1cBGpD0nUqCGtQqq78svw4fxA4ZgwBqV2yzIbgY1tY2o/edit#gid=1107419239");
var sSheet = sss.getSheetByName('SS Open Trades');
var sRng = sSheet.getRange("H3:J").getValues();
// Find last entry in columns H, I or J in the Open Trades sheet
for (var i = sRng.length-1;i>=0;i--){
var sLrIndex = i;
if (!sRng[i].every(function(c){return c == "";})){
break;
}
}
Logger.log("Last source row is %s", sLrIndex + 3);
// Check that the closed information has been added to all columns
// index 8 = 8th column = H = dateSold
// index 9 = 9th column = I = timeSold
// index 10 = 10th column = J = priceSold
if ((sSheet.getRange((sLrIndex + 3),8).getValue().length) == 0 ||
(sSheet.getRange((sLrIndex + 3),9).getValue().length) == 0 ||
(sSheet.getRange((sLrIndex + 3),10).getValue().length) == 0){
SpreadsheetApp.getUi().alert("You cannot process partially completed entries");
return;
}
// If there is a complete closed entry, copy the values to the variables
if (sLrIndex > 0)
{
coinBought = sSheet.getRange((sLrIndex + 3),2).getValue();
dateBought = toUtcString(sSheet.getRange((sLrIndex + 3),4).getValue());
timeBought = sSheet.getRange((sLrIndex + 3),5).getValue();
priceBought = sSheet.getRange((sLrIndex + 3),6).getValue();
dateSold = toUtcString(sSheet.getRange((sLrIndex + 3),8).getValue());
timeSold = sSheet.getRange((sLrIndex + 3),9).getValue();
priceSold = sSheet.getRange((sLrIndex + 3),10).getValue();
//Logger.log(priceSold);
}
else
{
SpreadsheetApp.getUi().alert("There are no completed entries to process");
return;
}
var tss = SpreadsheetApp.openByUrl("https://docs.google.com/spreadsheets/d/1umkTCr95FZUrZzv0e_ZDD9QZYiiYuH1fJohPYQzNE9Q/edit#gid=1644116137");
var tSheet = tss.getSheetByName('Trade Tracker');
var tRng = tSheet.getRange("B3:J").getValues();
// Create a new row at Row 5 on the Trades Tracker Sheet
tSheet.insertRows(5, 1);//shift all rows down by one from row 5
// Copy values from row 4 to row 5, including the formulae
var tRange = tSheet.getRange(4, 1, 1, 26);
tRange.copyTo(tSheet.getRange(5, 1, 1, 26), {contentsOnly:false});
// Populate row 4, first 7 table columns, from the variables
//Logger.log(coinBought);
tSheet.getRange(4,2).setValue(coinBought);
tSheet.getRange(4,3).setValue(dateBought);
tSheet.getRange(4,4).setValue(timeBought);
tSheet.getRange(4,5).setValue(priceBought);
tSheet.getRange(4,6).setValue(dateSold);
tSheet.getRange(4,7).setValue(timeSold);
tSheet.getRange(4,8).setValue(priceSold);
// Format the cells with dates and times
tSheet.getRange(4,3).setNumberFormat("dd MMM yyyy"); // Short Date
tSheet.getRange(4,4).setNumberFormat("HH:mm"); // Short Time
tSheet.getRange(4,6).setNumberFormat("dd MMM yyyy"); // Short Date
tSheet.getRange(4,7).setNumberFormat("HH:mm"); // Short Time
// Sort the sheet by date/time closed
// Find last entry in Trade Tracker sheet
for (var i = tRng.length-1;i>=0;i--){
var tLrIndex = i;
if (!tRng[i].every(function(c){return c == "";})){
break;
}
}
Logger.log("Last target row is %s", tLrIndex + 4);
var tRange = tSheet.getRange(4, 2, tLrIndex + 4, 8)
tRange.sort([{column: 6, ascending: false}, {column: 7, ascending: false}]);
// On the source sheet, delete the row just copied and add another blank row at the bottom of the table
sSheet.deleteRows(sLrIndex + 3, 1);
var rowLast = sSheet.getLastRow();
sSheet.insertRowAfter(rowLast - 1);
// Copy values from the new last row to the new previous to last row, including the formulae
var sRange = sSheet.getRange(rowLast + 1, 1, 1, 10);
sRange.copyTo(sSheet.getRange(rowLast - 1, 1, 1, 10), {contentsOnly:false});
SpreadsheetApp.getUi().alert("One completed entry processed");
}
Is there a more elegant way to write this? I dont want to use a for loop
if i==1 || i==6 || i==11 || i==16 || i==21 || i==26 || i==31 || i==36
function
end
basically i is the index of a vector, after each fifth element of this vector (starting with the first ) a specific function is applied. i starts with 1 and it increments after the if statement and just if it equals these values of the if condition the if statement is valid
EDITTED FOR MATLAB CODE OF MODULO
output = mod(input, 5); //this will output 1 if it is 1, 5, 11, 16
//input is your 1, 5, 11, 16 etc
//output is the result of modulo. else it is 0, 2, 3, 4
if(output == 1)
[previous answer]
i forgot how to write this in matlab but with your values, put it this way.
if(number%5==1)
any input 1 or 6 or 11 or any else that can add 5 to it, you'll end up in 1. else it will return false
I am writing a function that has to find the position of a given number within numerical ranges, the range is a variable within the code, for now lets say the range is 4 so the ranges will look like the following:
[ 0-3 ]
[ 4-7 ]
[ 8-11 ]
[ 12-15 ]
[ 16-19 ]
[ 20-23 ]
[ 24-27 ]
What i would like to achieve is to find the range where a given number belongs to, in the quickest way possibly,as this operation is performed over million of events.
So what i have wrote so far, and it works fine, is the following:
public String findRange(int range,int number2bFound)
{
int base = 0;
if (number2bFound == 0)
number2bFound = 1.0;
int higher = 0;
while (base <= number2bFound)
{
higher = base + (range - 1);
if ((base <= number2bFound) && (higher >= number2bFound))
return base + "-" + higher;
base += range;
}
return null;
}
So as i said this works, but i am sure this can be done implemented more efficiently, by only using the value of number2bFonud and the range and excluding the very expensive loop.
if all the ranges have the same size and it start from 0, a simple division will do, additionally you can find the position in the sub-range with a modulo operation.
The procedure is simple, find the integer division of your number n against yours range size and that will give you in which sub-range it belong, to find the position inside the sub-range find the modulo of your number again against the range size
here is a example python
def find_position(n,size):
return (n//size, n%size)
with range of size 4
>>> test=[ [0,1,2,3], [4,5,6,7], [8,9,10,11], [12,13,14,15], [16,17,18,19], [20,21,22,23] ]
>>> find_position(6,4)
(1, 2)
>>> test[1][2]
6
>>> find_position(11,4)
(2, 3)
>>> test[2][3]
11
>>>
with range 5
>>> test=[ [0,1,2,3,4], [5,6,7,8,9] ,[10,11,12,13,14],[15,16,17,18,19], [20,21,22,23,24] ]
>>> find_position(11,5)
(2, 1)
>>> test[2][1]
11
>>>
The procedure is a follow, let Size be the size of each sub-range and n the number you wan to locate, then you only need to the be the take the number n you want to find.
translate that to java should be very simple, excuse if my is a little rusty but is something like this I think
public String findRange(int range,int number2bFound){
int sub_ran_pos, pos;
sub_ran_pos = (int) number2bFound/range; //or however the integer division is in java
pos = number2bFound % range; //or however the modulo operation it is in java
return sub_ran_pos + "-" + pos; //or the appropriate return type, for this
}
(I don't remember, but if java is 1-index then you need to add 1 to each number to get right result)
I am writing a macro that processes an excel with lots of data. One of the rows contains a date range like wkstartdate - wkenddate and I would like to use dateadd function to print next date range every week (like '27-01-14 - 02-02-14' in below case) but unable to do so.
'06-01-14 - 12-01-14'
'13-01-14 - 19-01-14'
'20-01-14 - 26-01-14'
I used below excerpt which fails:
Range("E" & Lastrow).Select
prwk = Split(ActiveCell.Value, "-")
'curr_wkstart = DateAdd("d", 7, prwk(1)) 'error as maybe prwk(1) isnt correct format
'curr_wkend = DateAdd("d", 7, prwk(2)) 'error
Range("E" & Lastrow + 1).Value = curr_wkstart & curr_wkend 'no result
For testing purpose I print, prwk(1) which is 20/01/14 in the above case, in a diff cell and add 7 days, which gives me 1/21/2020 instead of '27/01/14'. I also tried using Cdate function, but still error
Can you please advise??
I think what you want to use here are the Format and DateSerial functions. Here's how I came at it:
Function GetNextWeek(TheStartWeek)
a = Split(TheStartWeek, " - ")
b = Split(a(1), "-")
c = DateSerial(b(2), b(1), b(0)) + 1
d = c + 6
GetNextWeek = Format(c, "dd-mm-yy") & " - " & Format(d, "dd-mm-yy")
End Function
Sub Test()
Debug.Print GetNextWeek("13-01-14 - 19-01-14") 'Givs you "20-01-14 - 26-01-14"
End Sub
Hope this helps.
ISSUE
I am trying to convert a 8 digit number into a date while in an array. Examples of the entries are 12282009 or 12202007. There are other malformed entries in the field including dates entered as strings. I want the 8 digit number to be formatted as 12/28/09 or 12/20/07 respectively. I keep getting a type mismatch error on the third to last line below. How do I do this??
CODE
Dim del()
ReDim del(1 To importwsRowCount, 1 To 1)
del = Range("AH1:AH" & importwsRowCount).Value
Dim delChars As Long
Dim delType As String
For i = LBound(del, 1) To UBound(del, 1)
delChars = Len(del(i, 1)) 'Determine length of entry
If IsNumeric(del(i, 1)) = True Then 'Determine datatype of entry
delType = "Numeric"
del(i, 1) = Abs(del(i, 1))
Else
delType = "String"
del(i, 1) = UCase(del(i, 1))
End If
If delType = "Numeric" Then
If delChars = 8 Then
del(i, 1) = DateSerial((Right(del(i, 1), 4)), (Left(del(i, 1), 2)), (Mid(del(i, 1), 3, 2))) '<-- TYPE MISMATCH ERROR
End If
End If
ENTRY TEMPLATES
SEPT. 25, 20 (No year, no year! Delete.)
SEPT (No year, useless, delete.)
N/A (Rubbish! Deleted.)
LONG TIME AG (What moron thought this was a good idea, delete.)
JUNE 30, 200 (Apparently the field will only hold 12 characters, delete.)
CHARGED OFF (Useless, delete.)
94 DAYS (Take all characters preceding space and subtract from other field containing order date to obtain delinquent date.)
94 DPD (DPD in someones bright mind stands for Days Past Due I believe. Same as above.)
2008-7-15 12 (Not sure what additional number is, take all characters before space and transform.)
INVALID (Delete.)
BLANK (Do nothing.)
4/2/4/09 (Malformed date, delete.)
1/1/009 (Same as above.)
12282009 (Use nested LEFT and RIGHT and CONCATENATE with / in between.)
9202011 (Add leading zero, then same as above.)
92410 (Add leading zero, this will transform to 09/24/10)
41261 (Days since 31/12/1899, this will transform to 12/08/12)
1023 (Days since delinquent, subtract from ORDER DATE to get delinquent date.)
452 (Same as above.)
12 (Same as above.)
1432.84 (Monetary value, mistakenly entered by low IQ lackey. Delete.)
Right(Left(del(i, 1), 2), 6) is nonsensical.
The Left(del(i, 1), 2) part happens first and returns a 2-character string. If you then apply Right(..., 6) to that 2-character string you get an error.
The Mid function is needed here: Mid(del(i, 1), 3, 2)
Running the Abs function earlier changed the array entry from being a Variant with subtype String to being a Variant with subtype Double. This shouldn't necessarily affect the Left/Mid/Right functions but try:
del(i, 1) = CStr(del(i, 1))
del(i, 1) = DateSerial((Right(del(i, 1), 4)), (Left(del(i, 1), 2)), (Mid(del(i, 1), 3, 2)))
We need to identify what the actual value causing the error is so:
If delType = "Numeric" Then
If delChars = 8 Then
On Error Goto DateMismatchError
del(i, 1) = DateSerial((Right(del(i, 1), 4)), (Left(del(i, 1), 2)), (Mid(del(i, 1), 3, 2))) '<-- TYPE MISMATCH ERROR
On Error Goto 0
End If
End If
' at the end of your Sub or Function - I'm assuming Sub here
Exit Sub
DateMismatchError:
MsgBox "Date mismatch: error number " & Err.Number & ", " & Err.Description & _
" caused by data value: |" & del(i, 1) & "| at row " & i & ". Original data " & _
"value is |" & Range("AH" & i).Value2 & "|, displayed value is |" & _
Range("AH" & i).Text & "|, number format is |" & Range("AH" & i).NumberFormat & "|"
End Sub
You can use this shorter code to replace your array elements with formatted dates
It cuts down the amount of testing inside the loop to two IFs. If numeric test is run first - there is no point running a longer lenint test for strings that are not 8 characters
The string functions Left$, Mid$ etc are much quicker than their variant cousins Left, Mid etc
I have made a substituion for your importwsRowCount variable in the code below
Updated code to handle and dump results, now handles string tests and non-compliantnumbers as per barrowc comments
The code below puts the new dates into a second array, skipping the invalid dates
The second array is then dumped at `AI``
Sub ReCut2()
Dim del()
Dim X()
Dim lngCnt As Long
del = Range("AH1:Ah10").Value2
ReDim X(1 To UBound(del, 1), 1 To UBound(del, 2))
Dim delChars As Long
Dim delType As String
For lngCnt = LBound(del, 1) To UBound(del, 1)
If IsNumeric(del(lngCnt, 1)) Then
If Len(Int((del(lngCnt, 1)))) = 8 Then X(lngCnt, 1) = DateSerial(Right$(del(lngCnt, 1), 4), Left$(del(lngCnt, 1), 2), Mid$(del(lngCnt, 1), 3, 2))
End If
Next
[ai1].Resize(UBound(X, 1), UBound(X, 2)).Value2 = X
End Sub