How to process a table serially for count - sql

Question

I need to process through a list of technical skills one by one and get a count of the number of developers we have in 3 locations who have that skill. For example, car type = "Java". How many persons have this skill listed in their resume.
I have 2 tables:
Skills: contains a single column listing skills - "Java" for example
Resources: contains 4 columns, Resource-ID, Name, Location, and a long text field called "Resume" containing text of their resume.
If this were a single skill I would process the SQL something like below (SQL syntax not exact)
SELECT count FROM [Resources] WHERE ([Resources].[Resume] Like "SKILL-ID*");
I want to process the Skills table serially printing the "Skill" and the count in each location.
Help appreciated
I've only used Access as a DB for single record retrieval, never using table values as input to loop through a process. I suspect that this is a simple execution in MS Access.

Answer 1

Ok, so we have to process that table.
I would simple "send out" a row with Location and skill for each match. We could write some "messy" code to then group by, but that is what SQL is for!!!
We could quite easy keep/have/use/enjoy the results in code, but its better to send the results out to a working table. Then we can use what sql does best - group and count that data.
So, then, the code could be this:
Sub CountSkills()
' empty out our working report table
CurrentDb.Execute "DELETE * FROM ReportResult"
Dim rstSkills As DAO.Recordset
Dim rstResources As DAO.Recordset
Dim rstReportResult As DAO.Recordset
Dim strSQL As String
Set rstSkills = CurrentDb.OpenRecordset("Skills")
strSQL = "SELECT Location, Resume FROM Resources " & _
"ORDER BY Location"
Set rstResources = CurrentDb.OpenRecordset(strSQL)
Set rstReportResult = CurrentDb.OpenRecordset("ReportResult")
Do While rstResources.EOF = False
' now for each resource row, process skill set
rstSkills.MoveFirst
Do While rstSkills.EOF = False
If InStr(rstResources!Resume, rstSkills!Skill) > 0 Then
rstReportResult.AddNew
rstReportResult!Location = rstResources!Location
rstReportResult!Skill = rstSkills!Skill
rstReportResult.Update
End If
rstSkills.MoveNext
Loop
rstResources.MoveNext
Loop
End Sub
Now, the above will wind up with a table looking like this:
So, now we can query (and count) against above data.
So, this query would do the trick:
SELECT Location, Skill, Count(1) AS SkillCount
FROM ReportResult
GROUP BY Location, Skill
And now we get this:
And you can flip the above query to group by skil, then location if you wish.
so, at the most simple level?
We write out ONE row + location for every match, and then use SQL on that to group by and count.
We COULD write code to actually count up by location, but that VBA code would as noted be a bit messy, and just spitting out rows of location and skill means we can then group by skill count, skill location count, or location, skill counts just by using "simple" sql against that list of location and skill record list.
So, now use the report wizard on that query above, and we get something like this:
Of course it is simple to change around the above report, but you get the idea for such a simple task as you noted.

Answer 2

Summarizing count of developers by skill and location can be accomplished with SQL. It requires a dataset of all possible skill/location pairs. Consider this simple example:
Resources
ID
Name
Location
Resume
1
a
x
Java,Excel
2
b
x
Excel
3
c
y
Excel
4
d
z
VBA,Java
SELECT Skills.SkillName, Resources.Location,
Sum(Abs([Resume] Like '*' & [SkillName] & "*")) AS DevCt
FROM Skills, Resources
GROUP BY Skills.SkillName, Resources.Location;
SkillName
Location
DevCt
Excel
x
2
Excel
y
1
Excel
z
0
Java
x
1
Java
y
0
Java
z
1
VBA
x
0
VBA
y
0
VBA
z
1
This approach utilizes a Cartesian product of Skills and Resources tables to generate the data pairs. This type of query can perform slowly with large dataset. If it is too slow, saving the pairs to a table could improve performance. Otherwise, a VBA solution will be only recourse and would likely involve looping recordset object.
Regardless of approach, be aware results will be skewed if Resume has content like "excellence". Bad data output is pitfall of poor database design.

Answer 3

Solution:
Resource and Skill table were added into MS_Access
Step 1: Create a query that executes the below SQL to get a counts (here named "Step1Query"):
SELECT Skills.SkillName, Resources.Location,
Sum(Abs([Resume] Like '*' & [SkillName] & "*")) AS DevCt
FROM Skills, Resources
GROUP BY Skills.SkillName, Resources.Location;
Step 2: Create a second query that uses the Step 1 query as input. (you can do this via the wizard):
TRANSFORM Sum(Step1Query.DevCt) AS SumOfDevCt
SELECT Skills.SkillName, Resources.Location,
Sum(Abs([Resume] Like '*' & [SkillName] & "*")) AS DevCt
FROM Skills, Resources
GROUP BY Skills.SkillName, Resources.Location
PIVOT Step1Query_qry.[Location];
Result lists out a matrix form. Thanks all for your help.