I have a collection of excel spreadsheets that are formatted... less than ideally.
I'm testing out some solutions involving SQLBulkCopy and OleDB, but I'm a bit concerned about how to handle the format of this sheet.
I was considering writing a custom Insert statement, but would like to see if there may be some easier way to implement a heuristic.
Below is a sample of the data I will be parsing:
The highlighted columns are the ones I'll be loading into the two tables. One table will hold order #s, and the other table will hold all the lines below that order number.
Any suggestions on tackling this would be lovely. The excel sheets are hand entered, so some weird cases exist (one order number with multiple carriers, which imposes the question of whether I should treat the first row with the order number as a line in the database structure I designed.
I'm implementing this importer within VB.net, to my dismay, to avoid being looked at funny by my coworkers :).
One approach would be to save the worksheet to a text file (e.g., CSV) and then use AWK to split it at the empty row. Some examples are in this SO answer: Bash how to split file on empty line with awk
You could then import the CSV files directly into the database.
Amusingly , if I wrote anything in VB.NET I'd definitely get looked at funny by my coworkers
So I'd use a library called EPPlus to read the excel and not have to worry about converting it. How you do the blank line detection is an open question- checking that the Value of ten cells on the row is Nothing or Empty would suffice. Then take the next row as your parent, and proceed with subsequent rows as children until the next blank
Take a look at this answer for more info on how to detect blank rows in Excel- if you get stuck turning any of the c# into vb shoot us a question. Online converters exist because the two languages are the same thing under the hood
I'm trying to create a sub-table from another table of all the last name fields sorted A-Z which have a phone number field that isn't null. I could do this pretty easy with SQL, but I have no clue how to go about running a SQL query within Excel. I'm tempted to import the data into postgresql and just query it there, but that seems a little excessive.
For what I'm trying to do, the SQL query SELECT lastname, firstname, phonenumber WHERE phonenumber IS NOT NULL ORDER BY lastname would do the trick. It seems too simple for it to be something that Excel can't do natively. How can I run a SQL query like this from within Excel?
There are many fine ways to get this done, which others have already suggestioned. Following along the "get Excel data via SQL track", here are some pointers.
Excel has the "Data Connection Wizard" which allows you to import or link from another data source or even within the very same Excel file.
As part of Microsoft Office (and OS's) are two providers of interest: the old "Microsoft.Jet.OLEDB", and the latest "Microsoft.ACE.OLEDB". Look for them when setting up a connection (such as with the Data Connection Wizard).
Once connected to an Excel workbook, a worksheet or range is the equivalent of a table or view. The table name of a worksheet is the name of the worksheet with a dollar sign ("$") appended to it, and surrounded with square brackets ("[" and "]"); of a range, it is simply the name of the range. To specify an unnamed range of cells as your recordsource, append standard Excel row/column notation to the end of the sheet name in the square brackets.
The native SQL will (more or less be) the SQL of Microsoft Access. (In the past, it was called JET SQL; however Access SQL has evolved, and I believe JET is deprecated old tech.)
Example, reading a worksheet: SELECT * FROM [Sheet1$]
Example, reading a range: SELECT * FROM MyRange
Example, reading an unnamed range of cells: SELECT * FROM [Sheet1$A1:B10]
There are many many many books and web sites available to help you work through the particulars.
Further notes
By default, it is assumed that the first row of your Excel data source contains column headings that can be used as field names. If this is not the case, you must turn this setting off, or your first row of data "disappears" to be used as field names. This is done by adding the optional HDR= setting to the Extended Properties of the connection string. The default, which does not need to be specified, is HDR=Yes. If you do not have column headings, you need to specify HDR=No; the provider names your fields F1, F2, etc.
A caution about specifying worksheets: The provider assumes that your table of data begins with the upper-most, left-most, non-blank cell on the specified worksheet. In other words, your table of data can begin in Row 3, Column C without a problem. However, you cannot, for example, type a worksheet title above and to the left of the data in cell A1.
A caution about specifying ranges: When you specify a worksheet as your recordsource, the provider adds new records below existing records in the worksheet as space allows. When you specify a range (named or unnamed), Jet also adds new records below the existing records in the range as space allows. However, if you requery on the original range, the resulting recordset does not include the newly added records outside the range.
Data types (worth trying) for CREATE TABLE: Short, Long, Single, Double, Currency, DateTime, Bit, Byte, GUID, BigBinary, LongBinary, VarBinary, LongText, VarChar, Decimal.
Connecting to "old tech" Excel (files with the xls extention): Provider=Microsoft.Jet.OLEDB.4.0;Data Source=C:\MyFolder\MyWorkbook.xls;Extended Properties=Excel 8.0;. Use the Excel 5.0 source database type for Microsoft Excel 5.0 and 7.0 (95) workbooks and use the Excel 8.0 source database type for Microsoft Excel 8.0 (97), 9.0 (2000) and 10.0 (2002) workbooks.
Connecting to "latest" Excel (files with the xlsx file extension): Provider=Microsoft.ACE.OLEDB.12.0;Data Source=Excel2007file.xlsx;Extended Properties="Excel 12.0 Xml;HDR=YES;"
Treating data as text: IMEX setting treats all data as text. Provider=Microsoft.ACE.OLEDB.12.0;Data Source=Excel2007file.xlsx;Extended Properties="Excel 12.0 Xml;HDR=YES;IMEX=1";
(More details at http://www.connectionstrings.com/excel)
More information at http://msdn.microsoft.com/en-US/library/ms141683(v=sql.90).aspx, and at http://support.microsoft.com/kb/316934
Connecting to Excel via ADODB via VBA detailed at http://support.microsoft.com/kb/257819
Microsoft JET 4 details at http://support.microsoft.com/kb/275561
tl;dr; Excel does all of this natively - use filters and or tables
(http://office.microsoft.com/en-gb/excel-help/filter-data-in-an-excel-table-HA102840028.aspx)
You can open excel programatically through an oledb connection and execute SQL on the tables within the worksheet.
But you can do everything you are asking to do with no formulas just filters.
click anywhere within the data you are looking at
go to data on the ribbon bar
select "Filter" its about the middle and looks like a funnel
you will have arrows on the tight hand side of each cell in the the first row of your table now
click the arrow on phone number and de-select blanks (last option)
click the arrow on last name and select a-z ordering (top option)
have a play around.. some things to note:
you can select the filtered rows and pasty them somewhere else
in the status bar on the left you will see how many rows meet you filter criteria out of the total number of rows. (e.g. 308 of 313 records found)
you can filter by color in excel 2010 on wards
Sometimes i create calculated columns that give statuses or cleaned versions of data you can then filter or sort by theses too. (e.g. like the formulae in the other answers)
DO it with filters unless you are going to do it a lot or you want to automate importing data somewhere or something.. but for completeness:
A c# option:
OleDbConnection ExcelFile = new OleDbConnection( String.Format( "Provider=Microsoft.ACE.OLEDB.12.0;Data Source={0};Extended Properties=\"Excel 12.0;HDR=YES\"", filename));
ExcelFile.Open();
a handy place to start is to take a look at the schema as there may be more there than you think:
List<String> excelSheets = new List<string>();
// Add the sheet name to the string array.
foreach (DataRow row in dt.Rows) {
string temp = row["TABLE_NAME"].ToString();
if (temp[temp.Length - 1] == '$') {
excelSheets.Add(row["TABLE_NAME"].ToString());
}
}
then when you want to query a sheet:
OleDbDataAdapter da = new OleDbDataAdapter("select * from [" + sheet + "]", ExcelFile);
dt = new DataTable();
da.Fill(dt);
NOTE - Use Tables in excel!:
Excel has "tables" functionality that make data behave more like a table.. this gives you some great benefits but is not going to let you do every type of query.
http://office.microsoft.com/en-gb/excel-help/overview-of-excel-tables-HA010048546.aspx
For tabular data in excel this is my default.. first thing i do is click into the data then select "format as table" from the home section on the ribbon. this gives you filtering, and sorting by default and allows you to access the table and fields by name (e.g. table[fieldname] ) this also allows aggregate functions on columns e.g. max and average
Might I suggest giving QueryStorm a try - it's a plugin for Excel that makes it quite convenient to use SQL in Excel.
In the SQL scripts Excel tables are visible as if they were regular database tables.
All four SQL data operations are supported: select/update/insert/delete.
The engine that executes the queries is SQLite so you can use joins, common table expressions, window functions, etc... And you get the fancy stuff like code completion, auto-formatting, symbol tooltips etc...
It has a completely free community edition for use by individuals and small companies. If you're in a company that has more than 5 employees or more than $1M in yearly revenue, you'll need a paid license but you can use a free trial key for evaluation purposes.
This blog post describes the SQL functionality of the plugin in much more detail.
Disclaimer: I'm the author.
You can do this natively as follows:
Select the table and use Excel to sort it on Last Name
Create a 2-row by 1-column advanced filter criteria, say in
E1 and E2, where E1 is empty and E2 contains the formula =C6=""
where C6 is the first data cell of the phone number column.
Select the table and use advanced filter, copy to a range, using
the criteria range in E1:E2 and specify where you want to copy the
output to
If you want to do this programmatically I suggest you use the Macro Recorder to record the above steps and look at the code.
The accepted answers here are old technology and shouldn't be attempted.
Back when this question was written, Power Query wasn't a well known option and wasn't available unless you were on the latest version of Office and installed it as a separate Add-in.
Now, Power Query is included in Excel and used by default to get data. It is the right way to do this. It is simple, fast and effective.
Here is the answer to the question in Power Query. Search on "getting started with Power Query" if you need help replicating this. Once you get started with Power Query, you'll see this is very basic and easy to do with the Advanced Editor:
let
Source = Excel.CurrentWorkbook(){[Name="Names"]}[Content],
#"Changed Type" = Table.TransformColumnTypes(Source,{{"lastname", type text}, {"firstname", type text}, {"phonenumber", type text}}),
#"Filtered Rows" = Table.SelectRows(#"Changed Type", each ([phonenumber] <> null)),
#"Removed Other Columns" = Table.SelectColumns(#"Filtered Rows",{"lastname", "firstname", "phonenumber"}),
#"Sorted Rows" = Table.Sort(#"Removed Other Columns",{{"lastname", Order.Ascending}})
in
#"Sorted Rows"
You can use SQL in Excel. It is only well hidden.
See this tutorial:
http://smallbusiness.chron.com/use-sql-statements-ms-excel-41193.html
If you need to do this once just follow Charles' descriptions, but it is also possible to do this with Excel formulas and helper columns in case you want to make the filter dynamic.
Lets assume you data is on the sheet DataSheet and starts in row 2 of the following columns:
A: lastname
B: firstname
C: phonenumber
You need two helper columns on this sheet.
D2: =if(A2 = "", 1, 0), this is the filter column, corresponding to your where condition
E2: =if(D2 <> 1, "", sumifs(D$2:D$1048576, A$2:A$1048576, "<"&A2) + sumifs(D$2:D2, A$2:A2, A2)), this corresponds to the order by
Copy down these formulas as far as your data goes.
On the sheet which should display your result create the following columns.
A: A sequence of numbers starting with 1 in row 2, this limits the total number of rows you can get (kind like a limit in sequel)
B2: =match(A2, DataSheet!$E$2:$E$1048576, 0), this is the row of the corresponding data
C2: =iferror(index(DataSheet!A$2:A$1048576, $B2), ""), this is the actual data or empty if no data exists
Copy down the formulas in B2 and C2 and copy-past column C to D and E.
If you have GDAL/OGR compiled with the against the Expat library, you can use the XLSX driver to read .xlsx files, and run SQL expressions from a command prompt. For example, from a osgeo4w shell in the same directory as the spreadsheet, use the ogrinfo utility:
ogrinfo -dialect sqlite -sql "SELECT name, count(*) FROM sheet1 GROUP BY name" Book1.xlsx
will run a SQLite query on sheet1, and output the query result in an unusual form:
INFO: Open of `Book1.xlsx'
using driver `XLSX' successful.
Layer name: SELECT
Geometry: None
Feature Count: 36
Layer SRS WKT:
(unknown)
name: String (0.0)
count(*): Integer (0.0)
OGRFeature(SELECT):0
name (String) = Red
count(*) (Integer) = 849
OGRFeature(SELECT):1
name (String) = Green
count(*) (Integer) = 265
...
Or run the same query using ogr2ogr to make a simple CSV file:
$ ogr2ogr -f CSV out.csv -dialect sqlite \
-sql "SELECT name, count(*) FROM sheet1 GROUP BY name" Book1.xlsx
$ cat out.csv
name,count(*)
Red,849
Green,265
...
To do similar with older .xls files, you would need the XLS driver, built against the FreeXL library, which is not really common (e.g. not from OSGeo4w).
You can experiment with the native DB driver for Excel in language/platform of your choice. In Java world, you can try with http://code.google.com/p/sqlsheet/ which provides a JDBC driver for working with Excel sheets directly. Similarly, you can get drivers for the DB technology for other platforms.
However, I can guarantee that you will soon hit a wall with the number of features these wrapper libraries provide. Better way will be to use Apache HSSF/POI or similar level of library but it will need more coding effort.
Microsoft Access and LibreOffice Base can open a spreadsheet as a source and run sql queries on it. That would be the easiest way to run all kinds of queries, and avoid the mess of running macros or writing code.
Excel also has autofilters and data sorting that will accomplish a lot of simple queries like your example. If you need help with those features, Google would be a better source for tutorials than me.
I might be misunderstanding me, but isn't this exactly what a pivot table does? Do you have the data in a table or just a filtered list? If its not a table make it one (ctrl+l) if it is, then simply activate any cell in the table and insert a pivot table on another sheet. Then Add the columns lastname, firstname, phonenumber to the rows section. Then Add Phone number to the filter section and filter out the null values. Now Sort like normal.